November 2018 Reply "subscribe" to be added, "un" to be dropped

Takeaways: Demonstrations of 5G phones but very few until summer 2019; Calix AXOS looks like a working SDN; 5G in Korea from Dec 1 and big in China; DT actively deploying an Edge Network;  Verizon 4G 1.45 gig; Huawei 50G PON; IoT is 4G, not 5G; Sckipio gig+; Say hello when we are in London“I've never seen a megahertz I don't like” Neville Ray
Houlin Zhao, just re-elected ITU Secretary-General, was sent down to the countryside along with tens of millions of Chinese teenagers as Mao’s system was fading. “It was very, very hard,” he once told me. “But I learned what life was like for a peasant. Today it informs everything I do.”     In Durban, Haolin quoted from Nelson Mandela’s electrifying speech to the ITU in 1995. “If more than half the world is denied access to the means of communication, the people of developing countries will not be fully part of the modern world. For in the 21st century, the capacity to communicate will almost certainly be a key human right. ... On Robben Island, when we were deprived of newspapers we searched the refuse bins for the discarded sheets of newspapers which warders had used to wrap their sandwiches.  We communicated with prisoners in other sections by gathering matchboxes thrown away by warders, concealing messages in false bottoms in the boxes and leaving them for other prisoners to find.”     Join me on the ITU 2030 Focus Group to define 6E. Let’s urge that standards provide for the most important enhancement of mobile broadband - access for all. or ------------“Mark Zuckerberg sincerely wants to connect the world,” a mutual friend tells me. Five reporters in the Times ripped into him, partly for playing dirty games lobbying. Mark is right saying, “This type of firm might be normal in Washington.” Imagine what five good reporters would find at AT&T. I sent to the Times some important datapoints on Facebook internationally.------------London, here we come. At Huawei’s wireless event, do say hello to the round fellow with a beard and the irrepressible Jennie Bourne. Thanks to Huawei for covering expenses.
  • Monday in London, we’ll have a small get together at 7 pm in our room at the Doubletree next to the Excel. Drinks and light refreshments. Email me if you are coming so we can send you the room number. 
  • Wednesday, from 5:30 to 7:30, I’m going to informally present the ideas behind 5G: The first three years in an apartment we rented in Kensington. Again, drinks and light refreshments. All welcome,
  • Invite me to present to your company in London or just drop by. No charge.
See you in London.5G phones from ZTE, Qualcomm (English newsbreak)Ted Rappaport promised there would be 5G phones for Xmas. Others agreed but they would be few, perhaps only a few dozen prototypes. ZTE is the first phone maker showing one. 20 companies have promised phones in 2019. Very few chips and hence very few phones will be available in the first half of the year.     A few days before, Swisscom connected the Qualcomm demo phone. A test network is live in Burgdorf, Lucerne, Zurich, Berne, Lausanne and Geneva. They want to go to 60 cities in 2019.     Many announcements to come, probably formally introduced at MWC Barcelona in February. The only one missing will be Apple, reports say. The 5G iPhone will be 2020 because Intel is a year late with the chips.*** ASSIA delivers reliable fast broadband to the home. Cloud based Management, diagnostics, and optimization serving over 100 million homes. (ad)Shanghai 10,000 5G cells in 2020; ?100,000 nationwide (English newsbreak)The official city plan calls for the "double gigabit" for all 9 million homes in Shanghai. FTTH is everywhere. 5G will deploy to 10,000 base stations by 2020, aiming for complete city coverage. The Chinese have chosen to emphasize Mid-band 5G for gigabit peak speeds. Users will usually connect at 100-400 megabits.      The city is also planning 100 cloud data centres with 16,000 racks. They are confident of over 100 "metropolitan IoT" applications. Shanghai, like Deutsche Telecom, is building further from the home but inside its own core network. For clarity, I call that an Edge/core network.      100,000 cells across China in 2020 is my wild guess. It certainly will be far more than last official estimate of 10,000 for the nation.

December 1 for 5G in Korea
The three Koreans, working with the government, had agreed to a March, 2019 official introduction. The date has been pulled up to December 1. In Seoul, SK is using a mid-band Samsung 3.5 GHz unit with 32 antennas. The unit is 1 meter in height and 23 centimeters in width, weighing 24 kilos. I believe the likely speed is one gig peak, 100-400 megabits typical. More, including a clear picture,

*** We have some ambitious plans for Internet Society New York. ISOC is free to join and choose the NY Chapter. Help appreciated. DaveCalix AXOS: "It's delivered, it's working, it's deploying."
Verizon's Lee Hicks is bringing Calix's ambitious new operating system into active deployment, a decision that would only be made after a careful review. Clayton Nash of CityFibre has adopted AXOS for a fully-funded fibre build to 5 million Brits.     This could be the first SDN system proving out in the field. I will be skeptical until we have more results from the field, but if Calix's AXOS delivers as promised, it will be interesting to nearly every carrier on the planet - telco and cable.    Calix also can offer the system completely as software. It will run on other company's hardware. They welcome other gear, from white box to sophisticated systems.

Pokémon GO first to DT "Edge network in the core"
Deutsche Telekom is building an "Edge infrastructure through decentralized cloudlets." Niantic Pokémon GO will be one of the first apps, with other AR/VR nets joining in. Controlling autonomous cars probably cannot work, but coordinating them and working with the connected car's onboard computers could yield remarkable results.     Cloud servers with ample processing power will be located in one of Deutsche Telekom's 17 facilities that connect all of Germany. What I'm calling an "Edge/core network" is designed for 20-25 milliseconds away, in parts of DT's backhaul system closer to the cloud than the consumer. More, including a video,

Verizon 4G 1.45 gigabits! Who needs 5G?
4G can be faster than 90% of 5G currently planned. Verizon achieved a peak speed of 1.45 gigabits using 40 megabits of licensed spectrum and a chunk of unlicensed Wi-Fi spectrum. LAA allows a telco to combine their licensed spectrum with any other available spectrum, such as the Wi-Fi bands. It usually adds 50% to 100% more capacity. Verizon is using mmWave for a true gigabit. Most other carriers are choosing low and mid-band. Mid-band is mostly 100-400 megabits, actual speeds to customers. 4G LTE might be 100-330 megabits, perhaps a quarter slower, with similar spectrum and antennas. LAA can add 200 megabits or so to LTE. T-Mobile LTE/LAA was independently tested at 500 megabits.*** Wireless Watch has a free trial offer, which will include the August 27th issue which includes the Verizon strategy report and an informed look at shared 3.5 GHz spectrum. (ad) WW is corporate-priced but delivers value for money. I asked Peter White how people can read that issue and he suggested just taking the trial offer. Caroline Gabriel's reporting from Facebook's TIP was very strong.

Huawei won't stop: 50G PON, 50G Ethernet, Revision 3, 5G & 20G microwave backhaul for small cells, low-cost 4G, remarkable Kirin 980
Tim Cook claims Apple's "A12 Bionic is the smartest, most powerful chip in a smartphone." He should have said, "One of the two most powerful chips in a smartphone." Huawei's Kirin 980 will probably beat it by some measures and do about as well as Apple in many others. Both are great chips, produced on the same TSMC assembly line. Hundreds or more likely thousands of first-rate engineers worked for years on both.    Huawei spends US$15 billion each year on research and now has the broadest product line. At the Geneva UBBF, I saw:
  • 50G Ethernet. 50G looks to be the primary backhaul in China. Huawei's unit featured state of the art HiSilicon network processors and some of Huawei's own optical units.
  • Gigabit mmWave units; four can be bonded for 20 gigabits. While the cliche is that 5G requires fiber backhaul, mmWave also has low latency. (In New York a few weeks later, Nokia showed a 10 gig unit.)
  • Ruralstar, Huawei's low cost, small, and low-powered cell radio. The growth today is coming from rural and less affluent consumers, especially in Africa and South Asia.
  • Amendment 3 for speeds over 1 gig. Swisscom is deploying it across the country.
"All current IoT applications work well with 4G, although I expect change in the future."
Dan Mondor, quoted, is CEO of Inseego, a US$200 million IoT specialist. Inseego sells a high-end Wi-Fi router to Verizon for the 5G network and works closely with Verizon. Even he couldn't think of any volume applications today that needed 5G. He expects that future applications will change that. Shane Eleniak is a Senior Vice President at Calix. His NG-PON2 and AXOS software are crucial to Verizon's 5G network. He confirmed to me that today's 4G can meet all present needs, unless there is an overall shortage of capacity.  Connected cars, remote surgery, and AR/VR/gaming are often cited as future demand-drivers. All are likely years away from having a large impact.

FCC small cell rules: "We don't believe we're going to see a material change in deployment."
Pai, O'Reilly, and the telco lobbyists proclaim the new FCC rules will result in far more 5G small cells deployed. They are wrong, as any conscientious reporter could have discovered simply by making half a dozen phone calls to the carriers to ask directly about the plans.     Some cities have ridiculous rules that should be changed, but the industry has long proven they can almost always be dealt with. The primary result will be lower charges to the telcos, perhaps as much as US$2 billion. I knew the promised construction boom was hogwash because the companies were telling Wall Street they would keep capital spending roughly flat. They won't be spending much on more small cells. CEO Jay Brown of Crown Castle, with the largest deployment of small cells in the U.S., "Doesn't believe we're going to see a material change." Verizon made a similar comment about their plans.*** Why Verizon thinks differently and what to do about it Verizon is going full speed ahead on millimeter wave. Nearly everyone else is doing the slower, cheaper mid-band 5G Low. What drives Verizon ahead. My new report for STL Partners. 5G: The first three years is readySckipio 212 Gig+ competing with 5G at SK Korea using 212 MHz of spectrum can deliver 1.5 gigabits+. That's enough to cover apartment buildings in Seoul and Seongnam, South Korea, as SK Broadband is doing. They are cascading two 24 port boxes from HFR, featuring the 24 port Sckipio chip. While Korea was one of the first to gigabit fibre, Sckipio reports two-thirds of Koreans do not have fibre all the way home. Most terminate the fibre in the basement, where is a natural upgrade for the original VDSL.     Sckipio is working to deliver 48 ports in a single unit.** Dean Bubley of Disruptive Analysis is presenting a small-group workshop in London December 4 on 5G, MVNOs & Network-Slicing. (ad) The always creative Dean is sure to have ideas you can use.

Engineer Nicki Palmer: "There's 5G and then there is Verizon 5G"
Verizon's millimetre wave 5G delivers a true gigabit to many. Around the world, most networks touting 5G will use mid-band with 60-80% less capacity. Verizon's latency will be particularly low because Lee Hicks' "One Verizon" plan will replace 200,000 current boxes with 20,000 all IP, state of the art gear. Hicks calculates they will save, "50% the first year." The new network will have fewer hops and reduced latency.       Verizon has not convinced people of the difference. They need to provide details of consumer benefits and direct comparisons. From bankers to regulators and from Congressmen to ordinary consumers, everything called "5G" is the same.

Ted's Armstrong Award & how 5G almost didn't  happen
Without Ted Rappaport, Verizon would not be turning on 5G today. In 2012, almost nobody believed millimeter wave was practical for consumer wireless. In 2013, NYU's     Ted Rappaport wrote Millimeter Wave Mobile Communications for 5G Cellular: It Will Work! based on thousands of tests his team did in the skyscrapers of New York.      Ted's results were much better than almost anyone else expected  He just won the Armstrong Medal from the Radio Club of America. He's particularly honored to win this because, "when I was a dirt poor college student, and newly married, this organization gave me a scholarship that helped me decide to stay for grad school instead of taking a job." The Awards banquet November 17 in New York will be, "An amazing night in general." Yes, I realize my headline is an exaggeration. Someone else would have discovered the properties of mmWave, probably years later. We wouldn't have it today.
  • Google Pixel’s new night vision mode is almost unbelievable. Try the sliders at
  • If you want good data, well presented, run, do not walk, to Incredible detail on everything from transit prices to total available bandwidth. Telegeography data has proven accurate for many years.
  • Tom Starr, who has led DSL standards and the Broadband Forum for more than two decades, writes, “I am excited to announce the publication of my second novel, ‘Fatal Entanglement.’  I think it is my best novel yet. It tells the adventure of a professor who enlists the help of a mysterious, genetically enhanced woman in a desperate struggle against foes who want to reshape the human species.”  Available as a paperback and ebook at Amazon
  • Will Goldman Sachs fund fibre across America? Goldman helped Greg Mesch of CityFibre raise US$3.3 billion to fibre five million in Britain. He now tells Iain Morris that similar funders would invest in the U.S. if they found fibre builders likely to succeed. Institutional investors like pension funds earn only about 3% on safe investments. They think fibre is a good bet if it returns 6-8%. That’s far less than the telocs expect in returns, so they are funding independents where telcos fear to tread.
  • Kevin Werbach has done important work at the FCC and University of Pennsylvania. He’s now written The Blockchain and the New Architecture of Trust. Kevin writes, “Blockchains and cryptocurrencies are the most exciting technologies I’ve encountered since the internet itself. They are also terribly overhyped, immature, confusing, and in the wrong hands, dangerous. It all comes back to trust. $19 at Amazon I’ve seen far more ridiculous hype than genuine uses, but some exist dave
  • Building a 5G backhaul network? You’ll want to attend the Broadband Forum’s NG-PON2 Roadmap and Evolution – Examining the State and Potential of a Universal Platform for Residential, Business and Wireless/5G. It’s at the OFC Conference, San Diego on Thursday March 7. The call for papers is open until December 15. rheier at for more information.
  • I’ve been hearing numerous complaints that NFV isn’t delivering close to what people hope. Iain Morris at Light Reading writes, “SDN NFV World Congress would have left many regular attendees feeling as if they had just received the same wretched birthday present for the umpteenth year in a row. Notes of desperation and disappointment colored the keynote presentations. ... Telcos are a long way from where they expected to be, executives at the show acknowledged.” Morris has more depth than I do on NFV, so I refer you to his article.
  • Ericsson CEO Börje Ekholm warns of serious interoperability problems in 5G. He’s not necessarily an objective source: both Ericsson & Nokia are trying to beat back new competition by raising fear, uncertainty, and doubt. It will be interesting to see how they respond to my request for examples and data.
  • The U.S. tariff on China is failing if success is more manufacturing in the U.S. Some work is deserting China but it rarely is moving to the U.S. Inseego, a fast-growing IoT shop that also makes home routers, said, “Actions we are taking are well underway, including the move of contract manufacturing of products sold in the U.S. out of Mainland China. Our plan is to complete the move by the end of this year.” All countries are proud and resist blackmail. Every elementary school child in China learns about the “unequal treaties” imposed after the Opium Wars, including by the Americans. There is a massive effort underway to replace the few remaining high components from the U.S., such as RF modules.
  • Cisco is doing another large layoff. Boosts profits in the short run, hurts soon after.
Reply "subscribe" to be added, "un" to be dropped
Volume 18, #18 November 2018​
October 25 Takeaways: Verizon and now Telefonica Germany are doing gig mmWave 5G; Verizon customers confirm 900 meg; 65% can be served from existing towers, hence cheap by telco standards; Rest of the world choosing the slower 5G/4G; Very few 5G phones until mid-2019; 10G fibre now so cheap forget GPON; Goldman Sachs raised US$3.3 billion to fibre 5 five million Brits; Massive corruption fine for Ericsson but CEO untouched; Swisscom & nbn Australia chose More

Telefonica CTO Enrique Blanco's eyes lit up when I suggested he could pull far ahead with true gigabit millimeter wave 5G. Deutsche Telekom & Vodafone are going with mid-band "5G" at 3.5 GHz, with 60-80% less capacity. He carefully told me nothing of his company's plans but we had a good discussion of why Verizon is betting US$20B on millimeter wave.
   Cayetano Carbajo, CTO Germany. has now announced that is exactly what they are going to do. They will offer fixed wireless to 25% of Germany, about 10 million homes, beginning in 2019.      They are the first in Europe to commit to mmWave; everyone else in Europe is sticking to 5G/4G mid-band hybrids. 
   Verizon is doing 25% of the U.S. and is already serving customers with close to a gigabit. See Verizon 5G: "I'm getting speeds of 900+ mbps downstream 200+ upstream." Planned right, the cost is so low Verizon can cut capital spending.
  • Calix Connections in Las Vegas in a few days will feature Lee Hicks of Verizon. He is tearing out almost all the 200,000 boxes in Verizon’s network. His One Verizon network will replace them with a tenth as many, all unified and IP. He tells me the savings are “50% in the first year” and that “Verizon costs are coming down 40%/year.” Today’s networks can be almost unbelievably inexpensive to run.
  • Say hello to the round fellow with a beard and the irrepressible Jennie Bourne. Whisper a story to me.
  • Next up, Huawei’s Wireless event in London in November, always very strong. I’ll stay over Nov 22 & 23 and hope to catch up with many of you.
  • NGMN Vancouver on Nov 5-7 will be one of the best events of the year. Respected CTO Ibrahim Gedeon will host senior people from China Mobile, Vodafone, and a dozen more telcos. Do go if you can; It just wasn't practical for me.
Next analyst report for STL: The Next Three Years of 5G. I think it will be interesting. I also have presentations on Why fibre is exploding worldwide and Why mmWave? Dave

Almost no 5G phones in 1H 2019 because Apple has bought 75% of world capacity
5G phones will be in short supply until the fabs, especially TSMC, add more high-end capacity. (Probably second half of 2019.) 5G New Radio requires much more complicated chips. The phones are working, but can't be produced in volume. 
     Without the most advanced chip processes, speed and power requirements cannot be met. Only three companies in the world are manufacturing at that level: TSMC, Samsung, and Intel. Only TSMC is shipping this year. 
     Apple has already bought 75% of what TSMC can produce in the near term. (EE Times) Huawei has claimed 10% of TSMC's 7 nm, much of which will be used for their remarkable Kirin 980 advanced 4G.  Much more

*** ASSIA delivers reliable fast broadband to the home. Cloud based Management, diagnostics, and optimization serving over 100 million homes. (ad)

Gig 5G: Telefonica's plan to blow open the German market
CTO Enrique Blanco's eyes lit up when I suggested he could pull far ahead with true gigabit millimeter wave 5G. Deutsche Telekom & Vodafone are going with mid-band "5G" at 3.5 GHz, with 60-80% less capacity. Cayetano Carbajo, CTO Germany. has now announced that is exactly what they are going to do. They will offer fixed wireless to 25% of Germany, about 10 million homes, beginning in 2019. They are the first in Europe to commit to mmWave. 
    Planned right, mmWave cost is so low Verizon can cut capital spending.  Hans Vestberg at Verizon estimates US$250-400 per home passed, half the cost of fibre. Reaching 25% can be done merely by upgrading existing cell sites, relatively cheap.  
   Going quickly to mmWave is probably the right strategy for non-incumbents where unbundling is not practical. Look out, Deutsche Telekom and Vodafone. Much more

You, yes you, should help define the wireless future at ITU Focus Group 2030. I did.

ITU IMT 2020 defined 5G. FG 2030 is an open group, free to all, setting the stage for 6G. You can make a difference while learning a great deal. There's an active mailing list and ITU believes in remote participation. All you have to do is go to and find the mailing list information in the lower right. You may need to register for an ITU TIES guest account, but it's free. Ask me for advice and how to connect effectively. 

GPON is Dead! Long live 10G. The low price of 10G is inspiring the switch
For just $5 or $10 more, it's a no-brainer to go 10G and let GPON fade away. A world class telco tells me they will switch in 2019. Xavier Niel's Salt tells me customers love the 10G. Jeremy Chelot's Community Fibre in London, Inea in Poland, and Hong Kong Broadband are doing well. AT&T has supported an Open Source 10G ONT and done extensive trials. 
     Charter has 25 million cable customers, nearly all of whom can get a gigabit. Comcast has 30 million. That's why AT&T is building gigabit fibre at the fastest pace in the western world; cable is killing them in the old DSL areas. 
   Adtran told me 10G is a no-brainer. Huawei confirmed.The difference in cost is so low you can and should take it out of the marketing budget.

*** Why Verizon thinks differently and what to do about it Verizon is going full speed ahead on millimeter wave. Nearly everyone else is doing the slower, cheaper mid-band 5G Low. What drives Verizon ahead. My new report for STL Partners.

2 Days, US$25B gap between VZ (better networks) & AT&T (DirecTV, TimeWarner)
Verizon wireless pulled up earnings, AT&T media held back profits. Both cut back capital spending, once more proving the "High cost of 5G" is a myth. Meanwhile, the prices of wireless service in the U.S. are going up as competition weakens.
   By the end of the day, AT&T was down 8%, about US$20 billion. During the last year, Verizon is up 17% and AT&T down 13%. The difference is about US$60 billion. Craig Moffett's take is that AT&T (and Sprint) are in thrall to the bondholders while Verizon and T-Mobile are delivering growth because that's what the stock market demands. He speculates T might have to cut the dividend for the first time in 30 years. "AT&T is prioritizing margins and cash flow over revenue and subscriber growth. ...Virtually every single part of legacy AT&T is shrinking. There are holes in every part of the story."
   AT&T, as I previously reported, has backed off on the 5G mmWave.  More, including a dramatic chart

** STL Partners will be presenting at Total Telecom Congress in London on Tuesday 30 and Wednesday 31 October. Phil Laidler will cover Telco Edge Compute; Amy Cameron the The Telco AI Roadmap; and David Martin NFV/SDN: What are telcos really doing? (ad) Andrew Collinson, who is chairing several sessions, has an interesting overview of the changes needed at I've been working with the STL people and seen some of the material. Interesting stuff.

Verizon 5G: "I'm getting speeds of 900+ mbps downstream 200+ upstream."
As expected, about three times faster than the mid-band 5G. CFO Ellis says VZ is getting what they expected, often just short of a gigabit. We're (finally) getting feedback from users, always more reliable than the companies. Several have reported on the web speeds usually just a little less than a gig. 
     Latency is 10-25 ms, not 1 ms. Verizon & AT&T latency from the cell to the home is testing at 9-13 ms. That's of limited value unless the game or whatever you're accessing is stored locally, unlikely for years. The important latency figure is from you to where you are trying to reach, sometimes across the web. 1 ms works in the lab but I do not know any telco in the world committed to deploying it in the next few years. more

** Dean Bubley of Disruptive Analysis is presenting a small-group workshop in London December 4 on 5G, MVNOs & Network-Slicing (ad) The always creative Dean is sure to have ideas you can use.

20% of Britain getting fibre from Goldman Sachs supported CityFibre
People are buying speed, so CityFibre could win enough customers to decimate British Telecom.  Vodafone is the anchor tenant, which should give them enough scale to be in the game. They are budgeting US$3.2 billion for 5 million homes passed, $600-700/home. 
     Telus CEO Entwistle explains, "We see churn rates on fibre that are 25% lower than copper. And that's encouraging, 35% lower in HSIA and 15% lower on TV, 25% lower on average. We're seeing a reduction in repair volumes to the tune of 40%. We're seeing a nice improvement in revenue per home of close to 10%." 
     CityFibre network costs should be > 50% lower than British Telecom. Any well-designed new network is half as expensive as the current telco networks. Lee Hicks at Verizon is ripping out 200,000 pieces of equipment left over from the many networks Verizon once had. (Phone, wireless, broadband, business, etc.) His One Verizon plan will replace all of them with a single IP network with only about 20,000 boxes.
   I asked Hicks, "How much do you think you will save?" The answer, "At least half in the first year." much more, including cable

*** Wireless Watch has a free trial offer, which will include the August 27th issue which includes the Verizon strategy report and an informed look at shared 3.5 GHz spectrum. (ad) WW is corporate-priced but delivers value for money. I asked Peter White how people can read that issue and he suggested just taking the trial offer. Caroline Gabriel's reporting from Facebook's TIP was very strong.

Qualcomm: "65% 5G millimeter wave coverage from existing towers"
5G is not going to be an expensive network build, according to Verizon, NTT DOCOMO, France Telecom, and Qualcomm. The latest data come from Qualcomm. Qualcomm simulations, based on data from field trials, show existing towers can cover two-thirds of most cities.“Based on our extensive over-the-air testing and channel measurements, significant outdoor coverage  (> 65%) is possible utilizing actual existing LTE sites."
   Upgrading existing cells first drastically brings down costs. Verizon and I believe AT&T are choosing to do most mmWave from existing cells. By the second phase of the rollout, 2023 or 2024, CEO Hans Vestberg predicts equipment costs will be far less. He expects to cover "the entire country" by 2018. Small cells will continue to be built where needed for capacity but few added for 5G. 
     mmWaves are blocked by some walls and windows, while they pass unimpeded through others. For now, Verizon is using only outdoor antennas. They hope customers will soon be able to do indoor self-installs in many cases. More, including a Verizon illustration of a dense area where fewer cells are needed in 5G.

*** We have some ambitious plans for Internet Society New York. ISOC is free to join and choose the NY Chapter. Help appreciated. Dave

Swisscom betting on across the country
CTO Heinz Herren is publicly confident that - usually at a gigabit - is right for the company's future. "Now that we have the new generation of chips, the time is right," he said in the keynote at Huawei's UBBF in Geneva. The new chips can use 212 MHz of spectrum and deliver fast upstream and downstream. They will be deployed to most of the country other than the third already covered with FTTH.
   The challenge, of course, is Xavier Niel's Salt with a ten gig connection for 50 Swiss francs, US$52. It's available to 1.3 million homes, about a third of the country.  I doubt there is much practical difference. between, say, 200 meg and 10 gig. The experience worldwide is that people do want the higher speeds if the price is similar.
    Jack Zhu of Huawei confirmed the new Amendment 3 chips from Broadcom are shipping in volume and the issues with vectoring have been resolved.

AT&T drastically cuts back on 5G; mostly lower and slower
AT&T wireless will be distinctly inferior to Verizon for years. AT&T's Andrew Fuetsch said mmWave will only be used in hotspots. Most of the "5G" network will be low and mid-band.       Over strong objections of their technical staff, AT&T has decided not to build millimeter wave 5G widely. Instead, they will use lower spectrum bands with 70% to 90% less capacity. 
    Today's mmWave 5G gear is designed for 28 GHz; the European 26 GHz is not supported by the new Qualcomm modem, although that will probably be remedied soon. No phone is likely to support AT&T's 39 GHz unless specially ordered.
   The soon to be auctioned 24 GHz is also unsupported. That means it is unlikely to be used for years. Pai will declare it a great victory, but the government would collect ?billions more if the auction were delayed. More

In the last two weeks, Mike Dano has reports Verizon is backing off from the plan to bring millimeter wave to most of the country and will use some 3.5 like AT&T. In addition, Verizon CFO Ellis has said they don’t expect “material earnings” from 5G until 2020, i.e. they will go slow in 2019.

Verizon CEO Hans Vestberg unlikely to go to jail despite US$B in Ericsson fine
Vestberg is a good guy doing a good job as CEO of Verizon. Ericsson made large payoffs when he was CFO (2007-2009) and CEO (2010-2016) of the Swedish company.  The company said the fines were likely "material," which for a US$24 billion (sales) company sounds like billions. 50 people have been fired. It is highly unlikely that Vestberg, a highly involved manager, didn't know what was going on. People at Vestberg’s level rarely go to jail. VZ is unlikely to fire him.
     The financial news from Ericsson showed a very welcome small profit. US$6 billion in sales was up 1% after currency adjustments and 9% without adjustments. U.S. sales were excellent, presumably especially at Verizon. 5G now works and carriers are buying. Networks are up slightly, services still awful. Huawei's US$15 billion research budget dwarfs all others, but Ericsson has increased spending on 5G. An analyst I respect, Carolyn Gabriel, thinks E is actually a little ahead. But it's very early.

"There is a clear and present danger of an imminent and great big famine engulfing Yemen."
The U.N. humanitarian chief Mark Lowcock told the Security Council that this famine would be "much bigger than anything any professional in this field has seen during their working lives." (AP) Yemen has/had an active chapter of the Internet Society but otherwise this is not a telecom story.

  • Thanks to Matt Wall for including me in his BBC story, Smart cities: 'A cyber-attack could stop the country.' If Facebook and Google can’t keep out hackers, how will any of us protect the 20 things we connect?
  • Update: Component shortages are continuing, per Adtran and Nokia. Suri pointing to IP routing in particular. Always watch for over-ordering that leads to excess capacity when caught up.
  • Mike Dano of Fierce out-reported me on Vestberg’s cutback of Verizon’s 5G buildout. Lowell McAdam had promised gigabit mmWave across the country. The October 1st announcement instead spoke of speeds of 300 megabits to a gigabit. I assumed Verizon was just being conservative, especially because they intend to have customers do self-install. I was wrong; Verizon last week spoke of using more 3.5 GHz, 70-90% slower. Mike was on target. His charts on U.S. spectrum at are very revealing. They show that Sprint has more spectrum per user below 2100 MHz than AT&T and about as much as T-Mobile. Added to Sprint’s massive holdings at 2500, they are second only to Verizon.
  • 120,000 Indians have lost their jobs as five of the eight largest telcos have disappeared in the last two years. The bloodletting is continuing. That’s the flip side of the incredible success of Reliance Jio.
  • Adtran had a profitable quarter and other good news. It is a solid company respected by customers. Based on the information publicly available, the recent stock price drop makes no sense.
Reply "subscribe" to be added, "un" to be dropped
Volume 18, #17 October 2018 
August 19

"Truth isn't truth." Rudy Giuliani speaking for Trump. 

Brett Feldman of Goldman Sachs stunned people with a prediction that Verizon would have 7,900,000 gigabit wireless homes in 5 years. 30M will be covered by gigabit mmWave, a quarter of the U.S. The cost will be so low capex will be little changed. It will raise profits significantly.
     Verizon CEO Hans Vestberg confirmed mmWave 5G capex will be about the same as 4G was. Verizon has done more testing of 5G than anyone else on earth. Vestberg is far more likely to be right than the many claiming brutal costs, no matter how distinguished they are or how many reports they pay for.
     Verizon's new data has induced a worldclass telco to re-run their numbers for 5G. They probably won't start a large mmWave build but I expect they will decide to aggressively test and train to be ready. McAdam of Verizon intends to prove mmWave will be a huge competitive advantage and others may follow.
Fiber to the home is also exploding internationally. Oi in Brazil just came back from insolvency and is completing a million homes passed. Spain and France have 40M homes passed. China has 328M connected. 
China Unicom, with 900K 4G base stations, may be taken over by China Telecom. If the rumors are true, it would be the most substantial deal in telecom history.
Say hello to this round fellow with a beard at Adtran in Huntsville this week or Huawei/ITU in Geneva two weeks later. Slip me a news story.
Do follow @closertwit, part of Analysis Branch. I hope to make it the most useful newsfeed in telecom. Also, we have some ambitious plans for Internet Society New York. ISOC is free to join and choose the NY Chapter. Help appreciated.

Goldman: 8 million Verizon 5G fixed mmWave by 2023
30M homes servable, 35% penetration in five years, significantly raising earnings. Brett Feldman made a gutsy call today. Brett has been accurate in the past about Verizon plans; he's the only other person who in Spring 2017 projected 30M. Brett's assumptions are. 
     Verizon will add 5G millimeter wave to all their existing cell sites. (I estimate 60,000-70,000.) This brings about 30M homes out of incumbent territory within 2,000 feet. He expects that to be mostly completed by 2022.
      Verizon has demonstrated a gigabit at 2,000 feet, although I believe VZ  has provided no data on what percent of customers within 2,000 feet will actually be gigabit capable.
      Based on a large rollout in 2019 going forward and matching the 35% to 40% take rate for Fios, Brett calculates a likely 7.9M subscribers for 2023. Verizon today has only 7M broadband customers. Feldman expects Verizon will not build many new cells in this period because they can reach so many from current sites. He sees little increase in capex.

Verizon CEO Vestberg: Being #1 in 5G will not raise capex
Almost everybody thinks 5G mmWave will be very expensive, require vast numbers of small cells, and may not pay off.  That's all nonsense, according to Verizon CEO Hans Vestberg. Here's what he said on CNBC:
  • 5G is the same cost as 4G was.
  • Verizon is the first in the world. It's going to take some time to cover everyone but not a decade.
  • We designed 5G to make everything cordless. Who needs landlines?
  • Spectrum is just one piece of the work. (Elsewhere, he's said Verizon doesn't need more spectrum. They have 800 MHZ at 28 GHz.)
Stephane Richard at Orange/FT agrees, "I just want to remind everyone that our view is that 5G will not mean significant additional CapEx for us in the coming years. This is what we have constantly said and explained to the market, and this is still our view."

*** Do follow @closertwit, a useful newsfeed.

Samsung: 2018 complete 5G chipset
Second source guarantees phones in 2018, but probably very few. Qualcomm's similar X50 chip is in the hands of testers, as is Qualcomm's remarkably small complete radio frequency front end. Samsung's new Exynos 5100 does 3.5 GHz and millimeter wave, with the expected bells and whistles.
      The surprise in this announcement is a complete chipset, including "radio frequency IC (RFIC), Envelope Tracking (ET) and power management IC (PMIC) solutions," I've asked for more details. They promise chips in 2018, but don't specify when volume shipments will begin. The rumors in the industry range from early to late 2019. more, including the full pr

First independent 5G 3.5 GHz results: Hundreds of megabits, not gigabits
Pal Zarandy tested Elisa's "world's first" 5G network from inside and outside. 5G at 3.5 GHz and Massive MIMO performs about the same as a good 4G network. "Gigabit LTE" - with tested speeds in the low or mid hundreds of megabits - has been shipping since 2016. Adding a minor software tweak, NR, and calling it "5G" did little to improve performance. (The battle to limit the term "5G" to true high-speed millimeter waves has been lost, unfortunately.) This is the dirty secret of "5G." Almost all the claimed new uses can be met with 4G. 
     The Elisa results are important because they confirm that Massive MIMO makes 3.5 GHz spectrum practical. Pal writes, "The key: 'Compared to gigabit LTE, the game changer features of the 3.5 GHz band won’t be single user speeds but instead the aggregate mobile data network capacity."
      Between 3.5 GHz and 4.2 GHz is enough spectrum to roughly double the capacity of today's networks.

*** Why Verizon thinks differently and what to do about it Verizon is going full speed ahead on millimeter wave. Nearly everyone else is doing the slower, cheaper mid-band 5G Low. What drives Verizon ahead. My new report for STL Partners.

Fiber to the Home near-explosive growth
20% growth is common. Telefónica Spain has passed twenty million premises - over 70% - and continues at two million a year. Telefónica Brazil is going from seven million in 2018 to ten million in 2020. China has 328 million connected and added 5M in the month of June. AT&T on July 26 announced 5M more in the next 12 months. 
    Telus CEO Entwistle explains, "We see churn rate on fiber that are 25% lower than copper. And that's encouraging, 35% lower in HSIA and 15% lower on TV, 25% lower on average. We're seeing a reduction in repair volumes to the tune of 40%. We're seeing a nice improvement in revenue per home of close to 10%,"
     What's going on? We're doing the edits on a report for the clients of STL Partners. The working title is Fiber to the Home near-explosive growth. I use the strategies of 15 major companies to frame the analysis. 
     Credit Suisse believes that for an incumbent, "The cost of building fibre is less than the cost of not building fibre."

AT&T fiber run rate going from 3M to 5M/year
John Donovan told investors they are now marketing fiber to 9M locations and he expects "that this time next year to reach 14M locations. ... We're doing very well in our fiber markets, including a 246,000 net increase in subs on our fiber network in the second quarter. ... Broadband penetration in the fiber footprint continues to be significantly higher than in AT&T's non-fiber footprint and is nearly 50% in locations marketed to for more than 30 months." 5M fiber homes passed in one year would be the fastest ever achieved in the West

"Worldwide shortage of components"
Carl Russo of Calix and Earl Lum confirm reports from Nokia & Joe Madden. Russo on the call said
"The caution on the revenue side is an extension of what we struggled a little bit within Q2. But we don't see that challenge going away and we're going to make some additional investments in inventory to deal with it, but there is a worldwide shortage of components that run straight through our product sets. Especially those product sets that are more subscriber edge premises oriented."

British Telecom replaces 3M with 3M fiber
Clive Seeley likes fiber and so does the British government. They are hiring 3500 engineers to do the installs. They've now cut 3M lines from their world-leading rollout and will go to fiber. BT, DT, and Verizon have been holding back from fiber to the home. BT & DT have changed their mind. (Verizon is using mmWave.) Iain Morris heard this from BT:
     "Full fiber's our priority and well accelerated our investment plans to reflect that ... Our engineers are already building FTTP to 10,000 premises each week and we're on track to reach 3 million homes and businesses by the end of 2020. "

Germany confirms 35b at "up to 250 megabits" to be available to 10M by yearend
Years late, DT is now satisfied 35G VDSL systems are working well enough to start selling it to customers. In a rare move, DT is also promising a minimum speed of 100 megabits. 
     35b VDSL is easy to explain. It the same thing as the popular 17b VDSL but uses 35 MHz of spectrum. Alcatel, Broadcom, and others thought the small change meant it would be available in 2016. It's much slower than, but should work well with DT's existing VDSL. Broadcom is the only chip supplier, something that scares away potential buyers, especially in a time of U.S. blockades.
     If the folks writing DT press releases have it right, DT will offer these speeds to ~30M households. 20% of Germany still has some of the worst Internet in Europe, despite promises and subsidies.

1M Ports of ship, $150M annual run rate
Alam Tamboli of Dell'Oro reports ~891,500 ports shipped by the end of Q1, which implies over 1M in Q2. The 2018 year over year growth looks to be 80%.  He also expects an annual run rate of over $150M/year by the first quarter of 2019, plus the CPEs. Adtran also has encouraging news. Australia is ready to take substantial volumes. 
    Dell'Oro shares with me their actual reports on the understanding I mostly use them for background and limit how much I publish.  Tamboli and Pongratz are consistently on target.

Analysis Branch 
Do follow  @closertwit, a useful newsfeed.

18 Whether to merge China Telecom & China Unicom will be high on the agenda of Vice Premier Liu He and STLG. 1 network is cheaper than 2, 2 networks are cheaper than 3. Competition can be costly.^ Everyone wants FTTH. Oi in Brazil, recently insolvent, is rushing past 1M homes.

16 Broadcom founder Henry Nicholas was arrested for drug trafficking. Painful to see.^Verizon CEO Vestberg says 5G costs will not be high but close to 4G costs and no more investment^Nvidia's new Turing design can do teraflops and motion picture quality graphics in real time.

13 Turkey is proposing a boycott of iPhones in response to American tariffs. Tim Cook is desperately afraid of same in China. He's buying Chinese chips and screens, investing heavily and begging Trump to help.

5G is real: China will see a "tidal wave" of 5G in 2019 to beat the U.S. (Madden) ^^ Prof Rappaport promises 5G phones by Xmas

Big wireless story Verizon will have 8M gigabit mmWave fixed by 2023, per Goldman. ^ Madden: political forces in China are driving a huge nationwide 5G low-band deployment tidal wave now . 

12 Verizon will have 8M gigabit mmWave fixed by 2023, per Goldman. ^ Government-backed vendor financing back Korea did $500M for Samsung Reliance Jio for 10 years. Canada, Belgium, and Finland are financing AT&T for Nokia. Error: Windstream up 20%, not 2%

9 BT replaces planned 3M with FTTH ^^ Deutsche Telecom turns on 6M VDSL 35b. 3/4ths get 175 megabits down. 10M year end ^^ DT cuts deal with Stuttgart for 500 million euros subsidy for FTTH

8 AT&T spends billion to take over Otter Media, now even more an "entertainment company." $130B spent in DirecTV & TW more than half AT&T's 236B market cap. ^^ Windstream jumped 2%% on lousy earnings. Market cap of $200M still implies edge of bankruptcy.

Other recent stories: August 6 Reply "subscribe" to be added, "un" to be dropped

4G or "5G will deliver a megabit or service for about 1/10th of what 4G does today." Lowell McAdam, Verizon CEO

Ted Rappaport predicts, "We'll have millimeter wave phones by Xmas 2018" But probably only "a few hundred," another source tells me.  
    Even more astounding, if confirmed, is Joe Madden's report that in China early in 2019, "political forces" will unleash a "huge nationwide 5G deployment at incredible speed" to pull far ahead of the Americans.
     Ericsson is shipping commercial software; Pal Zarandy connected on the commercial network of Elisa in Finland; Korea, Spain & England auctioned 3.5 GHz spectrum, netting from $0.05 per MHz per capita to $0.20; and the U.S. is seriously considering the Intelsat/Verizon giveaway of over $10B at 3.7 GHz.
     As you can see below, availability estimates differ, No one is sure whether there will be enough phones until about 2020. 
Most 5G, including China's, will be 3.5 GHz 5G Low, 70%-90% slower than 5G High (millimeter wave). Verizon is the only one committed to true gigabit 5G High in volume. AT&T and Korea are maybes, China and Japan probably mostly lower frequencies. 
    Why is Verizon spending about $20B to cover about 30M homes? Trials found mmWave goes much further than predicted, bringing down the cost. Lowell Mcadam is confident the much greater capacity (~3X) will win millions more customer. Much more in my report for STL Partners. 5G Why Verizon thinks differently and what to do about it More at end or
Fiber to the home is growing explosively. I didn't believe until I did the research. AT&T FTTH is going to 5M/year, up from 3M/year. Spain has 20M passed. Orange has 22M. Enel Open Fiber is funded at $10B to cover Italy. China has 328,000,000 connections, adding 5M more in June, Mukesh Ambani is just starting but is confident of 50M connections in India.
  20% growth rates are common. More next issue or in the second report I just finished for clients of STL Partners
215M 4G subscribers at Reliance Jio India is the second great story. It only took two years. The Indian government BharatNet has connected fiber to 300,000 villages and is on track for 300.000 more the next 12 months. India has passed the United States in Internet users and at this rate will have than the U.S. and Europe combined.
----------- isn't ready for its closeup but but check it out anyway. @closertwit intends to be the most useful telecom news feed and closer to the truth. Check it out and do follow.

Madden on China early 2019:  "Huge nationwide deployment at incredible speed"
"Political forces are driving. We ran the numbers, and they will surprise most of the industry. We had been limiting our forecast based on capacity issues in the supply chain, but with big recent investments in component capacity, we have bumped up our forecast for 2019."  Unconfirmed but plausible.
    The radios are available. Ericsson and Huawei are shipping production quality gear. Phones are uncertain but Qualcomm's small RF front end looks like a breakthrough. In March, I wrote 2M 5G Cells At 3.5 GHz for China Telecom.

***  Paul Budde and Andrew Collinson of STL Partners will do a webinar Wednesday August 8 5G: 'Just another G', yet a catalyst of change. Realistic deployment & uptake timelines •Role of enterprise/private networks •Fit with NFV, WiFi, edge computing & AI •The impact of geopolitics, spectrum regulation & lobbying No charge. Register *** Paul knows how to cut through the hype to what's really going on. Should be interesting.

5G mmWave phones for Xmas: Rappaport
"I told you 5G mmwave phones would be ready by Christmas," emailed Ted Rappaport, the "Prince of mmWave" and NYU Professor. A second source confirms the likelihood of "a few hundred" phones this year. Qualcomm has now announced they are sampling a small, highly integrated radio frequency and antenna module, the QTM052. I believe they will imminently announce they are sampling the X50 baseband modem and supporting a reference design.
     The early phones will be big, drain batteries rapidly, and very expensive. Few expect large volumes before 2020 and maybe later.

Ericsson: 5G is here! 5G mobile software shipping
Qualcomm expects to have enough mobile mmWave chips for a few hundred phones before Xmas. Unless Samsung or Huawei has a stealth project, those few hundred will probably be all the 5G phones in 2018. Buried in the bottom of this Ericsson-Verizon pr is "To prepare for 5G, Ericsson has released its first commercial 5G RAN software, allowing operators to turn on 5G in commercial networks when ready."
     None of the other four major vendors - Huawei, Nokia, Samsung, and ZTE - has announced their mobile version of the RAN software is commercial. The line between trials and early deployment isn't always clear. Ericsson's announcement does not imply they are ahead, but mobile 5G, at true gigabit speeds, is a great accomplishment for any company. "5G" is in "commercial production" at Elisa in Finland, Ooredoo in Qatar, Etisalat in UAE, & Saudi Telecom. They have all put out press releases they are "first in the world." (Elisa with Huawei.)

*** Do visit @closertwit, one of the most useful newsfeeds in telecom.

Smartphone: Apple falls to #4 behind BBK & Huawei
Who is BBK? The world's second largest smartphone vendor by units, BBK, owns Oppo, Vivo, and OnePlus. Oppo had 29.6M and Vivo 26M, totaling 55.6M. These giant brands are huge in Asia and are becoming very visible in Europe as well. 
     BBK operates as General Motors once did: Chevrolet, Cadillac, and Buick were run independently and competed with very different go-to-market strategies. Oppo and Vivo have large, independent research divisions as well. (Charts below.)
    Huawei emulates Henry Ford, who built steel mills and a rubber plantation to supply his automobiles. Huawei's HiSilicon is the largest chip company in China.

*** Why Verizon thinks differently and what to do about it Verizon is going full speed ahead on millimeter wave. Nearly everyone else is doing the slower, cheaper mid-band 5G Low. What drives Verizon ahead. My new report for STL Partners.

15% Wireless traffic growth at U.S. telcos !?
Even doubling the 15% growth rate leaves 2017 at far below expectations. Cisco's VNI, the best public forecast, has projected growth to fall to 30% in 2021, not 2017. I expect a rebound in 2018 but probably to less than the 40% productivity growth. The higher rates were an anomaly brought on by the introduction of smartphones. 
     Around the world, fixed and wireless, traffic per user has been falling.

Likely cities for Verizon millimeter wave
Atlanta, Detroit, Boston, Chicago, D.C., & Philadelphia/New Jersey - or a city nearby - appear to be close to ready to go commercial. Only Sacramento, Houston, and Los Angeles are announced. Plans could change so consider this a well-sourced rumor.
    Verizon has been doing trials in eleven cities. They are running fiber and beginning to deploy in literally hundreds of other places. Plans about which will turn on when can always change.

Verizon CFO: 5G low volume until late 2019 or 2020
Verizon CFO Matt Ellis in December 2017 said, "It just starts to be significant to our financials in the next 2 to 3 years." That would be late 2019 or 2020. Verizon is building a true gigabit millimeter wave network.
    In July 2018,  Skyworks CEO Kris Sennesael concurred. "We see revenue really being posted probably by 2020, maybe late '19. But 2020 is where I think kind of we translate around real revenues." Skyworks RF parts are in cellphones from all manufacturers. But see Madden's China report, above.

No 5G in Verizon "World's First 5G phone"
Add-on to come next year, at unknown price. Brilliant move by Verizon to announce a 5G phone that can't do 5G. It's the existing, not very impressive $480 Lenovo-Motorola Z3 and a promise of an add-on with a Qualcomm X50 modem. Modular, upgradeable phones like this have been around for a long time.

*** ASSIA and Incognito offer a joint solution to MSOs that tracks historical Wi-Fi performance for both TR-069-based and legacy CPE, continuously scans devices to determine optimal channels based on performance or speed, automates recommendations to improve QoE, and steers devices to optimal access points. (ad)

AT&T CFO skeptical on millimeter wave
AT&T is opening 12 cities with 5G mmWave in 2018, but hasn't announced anything past that. AT&T often announces number of cities but upgrades so few in each city the total is miniscule. Ask, what % of your subscribers will be able to connect and they will no comment. In this case, it's probably less than 1/2 of 1%.
    They will connect via a "puck," less demanding than mmWave in a phone. The technical people are gungho, as are some of the marketing people. "Whoever is first with 5G millimeter wave will win 10% to 20% of customers away from cable. We will be in a race to be first."
     CFO John Stephens isn't so sure.  Their fiber home is proving "Very inexpensive," and might be a better use of funds. They have raised the FTTH 12 month goal from 3M to 5M. The tone of this suggests they are still debating. 
     Verizon remains the only large mmWave build announced. AT&T and the Koreans are "maybes"

Massive MIMO: Vodafone Germany, Kolkata in Bengal, Parramatta in Australia & here in New York City
Massive MIMO produces a 2-7X capacity improvement, far more than the 15-50% of 5G NR software. Every engineer knows the 5G hype has become ridiculous, especially because Massive MIMO is a key component of 4G.
    Guido Weissbrich of Vodafone Germany confirms they have moved from trials to deployment. His colleagues are in production in Australia. Bharti in India is covering much of Kolkata, formerly known as Calcutta. Verizon is widely deployed in the U.S. but have released few details.
    All of these, I believe, are below 2.1 GHz and are FDD. FDD performance remains unproven and I'd welcome data. One vendor off the record guessed FDD Massive would only be a 1.5X improvement.  Karri Kuoppamaki of T-Mobile expects a 2-3X improvement. Marzetta tells me the solution is to change everything FDD to TDD, but that's not easily done.
   Sprint in the U.S. is going hard in TDD 2.5 GHz, including New York. They are reporting an average 4X improvement. That's less than Softbank Japan (up to 10X) and more than China Mobile (3X.) But performance varies so much by terrain comparisons are misleading.
    Almost all the work in 3.4-4.2 GHz is TDD Massive MIMO. It must be working well; 3.5 GHz spectrum is going for hundreds of millions and even billions.

  • Linley noticed I wrote Microtek rather than MediaTek. Thank you. If I lived in California, I would go to Linley Group chip conferences regularly. They are company-sponsored but he always brings in the best. His free newsletter is a must read for the geeks among us.
  •      In some issues, I used the link rather than for the article from the former FCC Chairman with data on Tmo’s 5G speeds. I also wrote "The AT&T CEO talked about the serious problems of T-Mobile" when I meant problems at AT&T. I disagree that AT&T faces an "existential threat," but he’s right about much of what he sees.


5G Why Verizon thinks differently and what to do about it is a new report I wrote for STL Partners and their clients.
STL Partners, a British consulting outfit I respect, commissioned me to ask why. That report is now out. If you're a client, download it here. If not, and corporate priced research is interesting to you, ask me to introduce you to one of the principals.
     It was fascinating work because the answers aren't obvious. Lowell McAdam's company is spending $20B to cover 30M+ homes in the first stage. The progress in low & mid-band, both "4G" and "5G," has been remarkable. In most territories, millimeter wave will not be necessary to meet expected demand.
     McAdam sees a little further. mmWave has 3-4X the capacity of low and mid-band. He sees an enormous marketing advantage: unlimited services, even less congestion, reputation as the best network. Verizon testing found mmWave rate/reach was twice what had been estimated. All prior cost estimates need revision.
     My take: even if mmWave doesn't fit in your current budget, telcos should expand trials and training to be ready as things change. The new cost estimates may be low enough to change your mind.
Follow @closertwit for all the above and more. Some reporting that might surprise you.
  • Hyperoptic raised £250m to run fiber in 50 cities. British Telecom is starting to feel pain.
  • Altice Europe dropped 16% after earnings revealed their low capex is hurting. Altice USA is doing well after a price increase.
  • ZTE spent $millions on lobbying D.C. It's the American Way. They can afford it. Chinese government banks have pumped in $10B.
  • Tsinghua Unicom has $174B for new fabs. Chinese General Processor Technologies is licensing CPU and other cores to take on ARM. Huawei is the first to a 7 nm mobile phone SOC, matching Apple and ahead of Qualcomm.
  • Craig Moffett sees Sprint insolvency despite good quarter.
  • Massive MIMO quadrupling capacity.  Massive MIMO 2-4 times more important than 5G Low
  • Nokia's joyous about a $3.5B possibility at T-Mobile but is making dangerous moves, fighting SDN & walking away from a contract at Verizon because the price was too high and another in China.
  • Vodafone CEO says data costs are down 90% in India after Jio entered. Everyone else thinks it's great news.
  • Almost all of Europe is clearing 3.4-3.8 GHz, allowing 100 MHz to each carrier. That is enough to double the capacity of today's networks. This once nearly untouched spectrum now is highly productive with Massive MIMO antennas.
  • Craig Moffett, one of the best on Wall Street, calculates that increased spectrum accounted for less than 1% of increased capacity in the U.S. Craig & Verizon's Hans Vestberg discuss how spectrum is always welcome but no longer is an important limit.
  • Verizon's Hans Vestberg just said 50% of their mid-band spectrum is unused. AT&T is similar. Except for India, spectrum is rarely the problem.
  • Verizon wants Pai to virtually give them ~$10B in spectrum around 3.7 GHz. Massimo MIMO makes it almost as valuable as low-band.
  • $8 buys 10-15 gigabytes on Australia's TPG, the new fourth carrier. It's free for 6 months, $10 (AUS) after that. Great argument for four, not three, mobile companies.
Reply "subscribe" to be added, "un" to be dropped
Volume 18, #14  August 6, 2018May 4 “Randall Stephenson portrayed the 140-year-old phone giant as being in an existential crisis." NY Times. He was under oath. 

John Legere of Tmo took a dangerous gamble Sunday; his pitch to D.C. was not supported by facts. They promised consumers a much faster Internet, sooner. I knew it was a crock because I had just written 5G NR Only 25% to 50% Faster, Not Truly a New Generation, with a definitive source.  His 5G NR 600 MHz is LTE with a software tweak, not much faster. 
       Sunday night, I had a 1200 word factsheet which went out to reporters. I often tilt at windmills, but this one is playing out. 
      By Tuesday, the NY TImes wrote “Implausible Promises.” They are too polite to say “lies” unless it’s Donald Trump. Half a dozen other reporters found Legere & Claure statements totally contradicted what they told the FCC. 
     Wednesday, former FCC Commissioner Wheeler picked up the 25%-50% figure, meaning everyone in D.C. will hear of it. (Thanks for the link
     Thursday, Sprint announced the "Best Financial Results in Company History."  Problem: the best chance to get the merger approved is proving Sprint goes broke without it.
     Wall Street puts the odds below 50-50 and the stock is down 20% on the week. The word is just getting out that it's almost impossible for T-Mobile launch a much faster Internet, sooner. Their obvious moves would require a time machine and break the Laws of Physics. If they can't deliver a faster Internet, sooner, all the promises about social benefits are void, Same for improving the U.S. position in 5G. Prices will be higher if the merger goes through than if it doesn't, according the best data
    T-Mobile's network is improving incredibly quickly. 500 megabit LTE/LAA is live in Manhattan, about the speed of their low-band "5G." They are on track for 30 cities in 2018 and the entire country by 2020 - including small rural roads. Taking over Sprint won't help much; they already have all the spectrum they really need for years, although more is always welcome.
    The New York Times may have killed the merger with “The Implausible Promises of a T-Mobile-Sprint Merger.” Maybe the lethal blow came at the BK 5G Conference last Wednesday, when the story broke low-band “5G” is little better than LTE FCC Chair Tom Wheeler used the 25% estimate 8 more articles linked at the end. 

Randall Stephenson’s contention that AT&T might die is a heckuva story. He was under oath in the TW antitrust trial. It seems wildly implausible. AT&T makes ~$20B/year, has $49B in cash, and claims $444B in assets, Doesn't sound like an "existential crisis." (NY Times.) But he has reasons for his fear and is cutting  70,000 jobs. Worth a click
80-year-old Wilbur Ross is mad to think killing ZTE will help U.S. technology companies. Apple stock has gone down $60B, three times as much as the total value of ZTE. Broadcom is dropped $8B & Qualcomm $6B. Intel would have to write off ~$10B in 5G if it can’t sell to 5 of the leading 7 phone makers: Huawei, Xiaomi, Oppo, Vivo &  ZTE. The Chinese fightback looks to be very strong. 
   Donald Trump explains Ross. “Wilbur has lost his step. Actually, he’s probably lost a lot of steps.”
Big stories to write: Pai is considering giving Verizon about $10B in 3.7 GHz spectrum for almost nothing.  The decision on the Intelsat spectrum isn't made yet, but the deal the companies want is ridiculous. * FDD Massive MIMO is working with a 2-3X improvement. * How Massive MIMO makes 2.4-4.2 GHz spectrum almost as useful as low band, virtually doubling usable spectrum. * Sharing 3.7-4.2 yields roughly twice the capacity as the planned auction. 

Quick notes: Wireless backhaul for wireless base stations is coming back into fashion. It’s now called IAB, Integrated Access Backhaul. ^^^ Excellent 30 pages by Andrew Schmitt on the state of the fiber art ^^^ Higher broadband speeds > 25 meg seem to have very little effect on usage of the net, Rob Kenny believes ^^^ 

One Trillion Dollars ($1,000,000,000,00) Supports China Fightback
“It is the compelling obligation for big companies to compete in core technology,” insists Jack Ma of Alibaba, market cap $451B, at the first Digital China Summit in Fuzhou. $457B Tencent chairman Pony Ma said "it is urgent for Chinese enterprises to make breakthroughs in the ownership of core technology." Xi Jinping sent a message to Fuzhou, calling for "Breakthroughs in core technologies."
    The Chinese will spend whatever is necessary to rapidly replace whatever ZTE can’t buy from the U.S. Microtek (Taiwan) is the #2 mobile chipmaker and Spreadtrum (China) #3. The Chinese government-owned telcos buy hundreds of millions of phones every year. It would be trivial to absorb the 15M ZTE has been selling in the U.S. 
     China has the talent and the funding. They’ve budgeted $100B just for memory chips.  The state-backed National Integrated Circuit Investment Fund just raised $32B. Shanghai has pledged $7.5B. In wireless standards, nothing goes through 3GPP without the approval of the Chinese. IEEE 802.11, which makes Wi-Fi standards, has a very strong Chinese presence. The ITU standards are dominated by C-J-K: China, Japan, Korea.
    The ZTE decision was an historic mistake, inspiring China to move even faster.

*** Sckipio GFAST. Ultrafast broadband without the sawdust. The way to get to ultrafast broadband without all the disruption. Delivering up to 1.5 Gbps today. (Ad) With cable going to a gig, anything less than is very risky. db

CEO: AT&T Faces An Existential Crisis. I Disagree.
AT&T makes ~$20B/year, has $49B in cash, and claims $444B in assets It's hard to believe CEO Randall Stephenson's sworn statement it faces an "existential crisis." (NY Times) 
     I've heard similar from others at the top of AT&T. In 2016, Randall feared, "In three years we’ll be managing decline." AT&T is working on ~70,000 job cuts, the largest cutback anywhere in the world since Elisha Gray and Alexander Graham Bell invented the telephone.
     Randall thinks Google and Facebook will take T's profits unless he remakes the company in their image. AT&T revenue was down $3B from 2016 to 2017. One analyst projects AT&T video revenues will go down from $31B in 2017 to $24B in 2023.
    They believe Google's datacenter costs are 50-80% less. T is turning the whole company around to a "web-centric" model faster than any other telco on Earth. They've embraced open source and cheap generic hardware. They plan to shut down 15% to 30% of their landlines and go wireless only.
    My take is they should can the TW deal. $85B is too much to pay for Time-Warner except as a disguised way to diversify from the prospects of the telco business.  Maybe Randall expects the telco decline will be much faster than consensus. He's been right when I and most others have been wrong.
    Time Warner earns ~$5B and Hollywood's business model is in trouble. The Chinese are suddenly making blockbusters. Detective Chinatown 2 and Operation Red Sea are #2 & #3 in worldwide box office this year. 
    AT&T, despite what the Times thinks, is not going to go out of existence

***ASSIA’s ClearView® Solution Brings Machine Learning to Customer Care Processes. The recommendation engine leverages machine learning to improve the customer care process. CloudCheck now supports Wi-Fi, GPON, DSL, & Cable. (ad) Amendment 3 Reaches Carriers
"We have A3 solutions in trials at large customers," Geoff Burke of Calix writes. Speeds are well over a gigabit. Calix has well over 100 customers, thousands of systems deployed, and tens of thousands of home gateways Analyst Alan Tamboli believes, "Once the higher frequencies (212 MHz) and higher port groupings (i.e., 48-port and 96-port) are successfully tested and trialed, this will change, especially at MDUs." 
   Calix uses Sckipio chips. Dynamic time allocation allows speeds to often be at 1 gigabit today. Success for Israeli Sckipio chips is good news for ZTE. They need to find non-U.S. suppliers for  Germany and Austria.
   Separately, Adtran announced AT&T is expanding in nbn in Australia will go in Q4.

*** London June 19-20 Connected Britain is the premier meeting place to discuss, learn and collaborate on the UK’s digital future. (Ad) The dynamic Sharon White of OFCOM keynotes; Everybody who is anybody attends. Point-Topic’s Oliver Johnson is a chair.

Goldman Sachs Puts $750M Into UK Fiber Boom
Goldman & Antin just bought CityFibre for $750M, a rich price. Fiber is back, bigtime. Pension funds and others want long term investments they believe will have a regular return. With the new investors, CityFibre plans 5M homes. They expect costs under $1,000/home. probably well under.
    Enel Open Fibre in Italy just collected $3.5B towards an $8B plan. AT&T has added $1B to raise the 2018 fiber build to 4M. Telefonica is over 15M homes passed, about 75% of Spain. Orange is not far behind in France & Spain. Iliad/Free is now up to 6M in France after a very slow start. Liberty Global is doing millions of lines. Bell Canada, Telus, and even Telefonica Brazil are doing millions more.
    TalkTalk, Britain's #4 broadband provider, announced a deal with Prudential M & G, another infrastructure fund. They plan three million lines for $2B. As far as I know, the deal isn't final yet and recent TalkTalk financial problems could derail it. More, including investment explanation,

Pai's 24% Solution For 5G Means Verizon + AT&T Own Almost All Prime mmWave
3/4ths of the U.S. won't even be available in the highly touted 28 GHz auction. There is essentially no spectrum available in the crucial 28 GHz spectrum band in the top 50 markets.
     Verizon controls 76% of the 28 GHz band in the top markets. Only 2% will be available for auction. Verizon has 46% of the 39 GHz band and AT&T 30%. 
   Stephen Wilkus calculates the 28 GHz auction will only cover 23.7% of the U.S. population. Much will be rural, inappropriate for the short reach of mmWave. 24 GHz will be auctioned right after 28 GHz finishes; but the only 24 GHz capable equipment I could find was a NASA funded research project. more, including how the Straight Path deal gives Verizon dominance

Correction, maybe: If the NR vs. LTE difference is 25% or more, I have an error. I’m waiting for firm data.

  • Are you going to The Hague for TNO in June, asked someone introducing a very interesting new product. It’s an excellent show I’d love to attend but I had to cut back on travel. The Mauritshuis in Den Haag has an incredible collection of Vermeers and Rembrandts. Don’t miss the Dutch coffee shops.
  • Om Malik sent the question, What is 5G? It's the right question. Two years ago, 5G was millimeter wave and maybe Massive MIMO. It was a big step up, mostly because there's enough empty spectrum for 400 MHz and 800 MHz channels. But once "5G" became incredibly hyped, the companies involved extended the term to something totally different: low and mid-band under 6 GHz. Great for marketing, of course.  The ITU is still holding out for IMT-2020 which peaks at 20 Gig. But the companies came together at 3GPP and gave the name 5G to just about everything they are building, The giant companies at 3GPP - now lead by China Mobile and Huawei - will probably get their way. Qualcomm, Nokia, and telcos around the world gave support, liking to market as 5G. Qualcomm led the vendors into a standard that will add over $5B a year to royalties and will take the biggest share.
  • It's good to put the spectrum to use - it can double the total of spectrum available. But T-Mobile & China Telecom are using the glitter to pretend their 600 MHz to 2.7 GHz is wonderful, T-Mobile wants to get rid of a competitor so they are huffing and puffing. China Telecom agreed with the government to build 2M 3.5 GHz small cells, pretty substantial but much cheaper than the mmWave originally planned.
  • 60 GHz is coming back, both for backhaul and home connections. Amitava Ghosh has important technical details Paulraj confirmed 60 GHz is looking good, at least for backhaul. Jennie has video to edit of Nokia’s remarkable 90 GHz ?256 element phased away. Antennas & array elements get smaller at higher frequencies so you can use more of them. Several people told me that will allow going to even higher frequencies.
  • Intel Capital and VC Sequoia led a $100M funding round for Horizon Robotics, a Chinese company doing advance AI including facial recognition. Selling the rope, anyone?
  • The only reporting I’ve seen bring important new information to the Sprint/Tmo debate is from Sheila Dang and Diane Bartz at Reuters. They note that the poorest Americans will be hit hardest. T-Mobile and Sprint lead the much lower priced prepaid market. They also are the support of most MVNOs, which serve many of the poor. Dang is recently out of Columbia Journalism but has been on target in the articles I’ve read.
  • Face recognition is working unbelievably well in China, resembling Big Brother. But that’s another story.


Too long. I didn't have time to make it shorter.

Here’s the nine part factcheck
On the effects of 4-3; On whether 5G will come faster; "Combining is the only way they can fight back against the two industry leaders;" On whether Sprint could survive without the merger; On whether Sprint could survive without the merger; On whether the T-Mobile 5G is a "seismic shift;" 5G is not 5G: Much will not be the truly high speeds of millimeter wave; Is even the real 5G, millimeter wave, revolutionary?; "5G offers more reliability than 4G or LTE"; Is 5g required for autonomous cars?' Is 5G needed for telemedicine or IoT?
Far too much more

5G NR Only 25% to 50% Faster, Not Truly a New Generation
The data that should kill the deal. They hate me in Seattle but truth is a defense for libel.

T-Mobile Watch: Chances Falling As Errors Reported, No Time Machine
The claims, the data, and more.

The Smoking Gun on 5G Speeds
Bell Labs chart shows you immediately how LTE & 5G are comparable

Professor Rappaport Explains Why T-Mobile 5G 600 MHz Ultimately Doesn't Make It.
A quick look at the technology reported above.

Will Times Reporting "Implausible Promises" Kill T-Mobile-Sprint?
With the Times warning, everyone now knows to check the all statements from Tmo. 

Wall Street's Best: TMO/Sprint Only Has 50% Chance: Brief
Moffett and others are skeptical.

Reply "subscribe" to be added, "un" to be dropped
Volume 18, #13  May 3, 2018

April 18

“My three biggest customers are marching to 7 and 5 nanometers.” Lip-Bu Tan. Moore’s Law ain’t dead yet.

Eric Xu, current Huawei Chairman, said consumers would find no “material difference between 4G & 5G.” Dozens of top engineers agree the "5G Revolution" is hokum, including  DT CTO Bruno Jacobfeuerborn, FT/Orange SVP Arnaud Vamparys, and BT CEO Gavin Patterson.
     Politicians and marketers pushed b_______. What’s coming for 5-10 years is good for telco capacity but otherwise disappointing. Unfortunately. 
     On the other hand, mmWave & Massive MIMO are improving at a ferocious rate, advancing telco capacity enormously. Massive MIMO has made it practical to use 3.5 GHz frequencies, a second remarkable booster. The "5G Revolution" is dead; Long Live the real 5G and advanced wireless!
U.S. attacks on Huawei and ZTE look to be an historic disaster.

China’s logical response is to invest to become independent of the U.S. They already have $100B on the table for memory chips and are rumored to add another $100B. Huawei/HiSilicon already holds their own against Qualcomm and everyone else in the world. 
    Zeng Xuezhong of Spreadtrum, world #3 mobile chipmaker, confirms: China will “speed up the chip R&D investment that is self-controlled and self-researched. We must never forget to make it bigger and stronger.”
     U.S. warships patrolling China’s seas inspire security fears. So does Trump’s threat to a country on China’s border."Military solutions are now fully in place, locked and loaded, [for] fire, fury and frankly power, the likes of which the world has never seen before.”
     China would be stupid not to have domestic capacity just in case. The Chinese leadership is not stupid.
The year's best wireless conference is next week. Ted Rappaport, Sandeep Rangan and colleagues bring the world's best engineers to the Brooklyn 5G Conference Summit. This year, Paulraj, Onoe, Fettweis, Katabi, and their peers are coming. Do not miss the webcast, .

No “Material Difference Between 5G & LTE”
Eric Xu, current Huawei Chairman, concludes "consumers would find no 'material difference between 5G & LTE'.” Louise Lucas and Nic Fildes,  Financial Times He added, "Since 4G is robust, we don’t see many use cases or applications we need to support with 5G.”
   Last May, I reported similar thoughts from Telefonica CTO Enrico Blanco. The politicians & marketers screamed "5G Revolution." The engineers knew better. 
   Andrus Anders and Roberto Viola at the EU, as well as Jessica Rosenworcel & Ajit Pai in the U.S., are still lying to themselves and too proud to face their errors.
   Latency:  Ericsson has promised LTE latency of 9 ms in 2018. AT&T's 5G latency is 9-11 ms.
   Speed: 4G LTE 2018 is hundreds of megabits, peaking over a gigabit. 90% of 5G on the way is midband, the same hundreds of megabits. Only 10-20% of the first few years will be millimeter wave, often a gigabit.
    Applications: 5G's main application around the world will be more capacity, a good thing for telcos. But connected cars are already on the road, using lidar & radar. The talk of remote surgery is totally ridiculous unless you believe in operating from the beach. IoT will rarely require speeds like 100's of megabits. Most actually is kilobits. VR experts tell me they are fine at 10-15 ms. More

*** ASSIA is proud to announce that Telefonica has deployed CloudCheck® software solution across Telefonica’s Latin American Networks. The CloudCheck solution is a highly scalable, big-data-analytics, and machine-learning platform that proactively detects and resolves subscriber Wi-Fi issues while automatically optimizing Wi-Fi performance for their subscribers. (Ad)

Australia Putting VDSL in Locations
Australia's nbn is running fiber close (FTTC/FTTdp) to 1.5M homes, the end of the driveway. could go well over a gigabit the short distance. G.mgfast could carry 4 gig the length of a very long driveway. (50 meters.) VDSL doesn't make sense unless most of your customers are 500 meters or more. 
    The difference in cost between VDSL 35b (200/40) & (750/750) works out to less than $2/month unless you are wildly overpaying. BT proved works by installing a million lines.
    Customers do want the higher speeds. Alan Breznick reports 10-20% of new customers at Mediacom are buying a gig, even though prices start at $125/month. Australians are ordering slow speeds (25-50 megabits) because unfortunate political decisions make better speeds far more expensive than costs would justify.
    Everything at nbn is intensely political.The national network has been at the center of politics for a decade. nbn is somewhere between $10B & $30B below breakeven, depending on your assumptions. It will need a direct or indirect government subsidy unless drastic action is taken. 
    The cost saving here is too small to make a difference.

*** Sckipio GFAST. Ultrafast broadband without the sawdust. The way to get to ultrafast broadband without all the disruption. Delivering up to 1.5 Gbps today. (Ad) With cable going to a gig, anything less than is very risky. db

Ich bin ein ZTE - China's Response
Hu Xijin, editor of government-run Global Times, wrote "We are all ZTE tonight," echoing Jack Kennedy to support the 80,000 workers of ZTE.  Apple is very, very scared the Chinese people will stop buying, as they did Samsung.
    The ZTE ruling portends that the United States government could shut down at will virtually any phone or telecom supplier in the world. Google may have to cut off Android apps, destroying ZTE's phone business, which ships 40,000,000 units a year.
    Qualcomm and Broadcom have committed to license all their patents in standards on "reasonable and non-discriminatory terms." The U.S. government has now ordered them otherwise. Disputes like this could destroy the entire standards process.
   It could go beyond the Chinese. For example, Samsung could be affected if the U.S. wanted to punish South Korea for making a separate peace with the North.

4 To 3 Wireless Mergers Doubled Relative Prices: Rewheel
Good data, suggestive if not definitive stats. Pal Zarandy at Rewheel finds "Gigabyte prices in 4 to 3 consolidated German and Austrian markets have fallen considerably behind the Netherlands and other 4-MNO European markets."
    Wall Street analysts are exceedingly skeptical the deal will go down and be approved. Top analyst Craig Moffett estimates the chances are 40%; a team at Goldman did not offer a figure but was even more pessimistic. Still, Sprint's stock jumped 18% in one day when WSJ reported discussions. Masa-san and Legere are ready to risk $billions that the deal will go through.
     Much more on Evidence-based analysis, Why Most Published Research Findings Are False by John Ioannidis, and worthless DC studies

Suddenly Samsung's Somebody; It's "5G Big 5" Now As Verizon & Sprint Go For Samsung
Their first large customer outside Korea was KDDI, Japan's #2. Samsung went on to win about a third of Sprint's U.S. LTE network, 13,000 base stations. Youngky Kim and team then won Reliance Jio, which became the fastest growing network on Earth. Jio now has 150,000 bases, about as many as Verizon and AT&T combined. Despite adding well over 100M paying customers since September 2016, Jio has consistently raised speeds and generally led the TRAI speed tests. 
    Dozens of Samsung engineers have written important papers. They are now appearing on most 5G shortlists in the parts of the world they are ready to service. 
    By February, Verizon had deployed 100's of 5G cells for testing and is committed to sighing the first customers by the end of 2018. (Both Verizon and AT&T have said they will be first. Pure pr.) Verizon is clearly building as fast as they can; I reported last spring the decision had been made to spend $20B to cover 25%-30% of the U.S.   
    The financials suggest Ericsson and Nokia may have no choice but to merge.

*** Brooklyn 5G Summit is an event not to miss. April 25-26 Sponsored by NYU Wirelesss and Nokia. IEEE Livestream. (psa) More than a dozen of the world's best wireless engineers are coming. See you there. Dave

Wireless Chips 2018 +1.5%: Apple, Samsung Struggling For Phone Customers
In the current market, Qualcomm has 40% while Apple-Samsung-HiSilicon have taken ~30% inhouse. "Smartphone application processors (AP) worldwide will grow 1.5% to 1.67 billion units in 2018," DIgitimes estimates. 
    Apple sales have been disappointing. The Chinese public is angry about ZTE & Huawei, with some talk of boycotting Apple. That happened to Samsung when the Koreans put U.S. THAAD missiles in place. The Samsung market share in China is now less than 1%.
    Consensus is smartphone sales in 2018 will be about flat. Some China Q1 sales are low enough to suggest a decrease in 2018. 
    What's interesting here is the market shares. As smaller phone vendors fall out, the share of self-produced chips rises. Apple, Samsung & HiSilicon are usually the three biggest phone vendors and predominate in the more profitable high end chips. More

New York's Truthtelling CTO Miguel Gamiño: "No Choice" But To Quit FCC Broadband
The FCC Broadband Committee recommendations are "Industry priorities without regard for a true public-private partnership. ... 75 percent of members represent large telecommunications and cable companies or interests aligned with those companies, ... These circumstances give me no choice but to step away."
    Miguel Gamiño Jr., Chief Technology Officer of the City of New York, wrote the letter below to FCC Chairman Ajit Pai. Like San Jose Mayor Sam Liccardo, he quit. He worked hard for 11 months but was overruled by the company people Pai put on the Broadband Deployment Advisory Committee. 
    Gamiño is refusing to accept a "Pre-packaged one-size-fits-all proposal that industry lobbyists have pushed nationwide rather than working in a cooperative fashion to find creative solutions to dynamic local issues."
     Pai's claim that getting rid of regulations on small cells will have a major impact on the deployment of 5G is almost certainly wrong. The telcos have already decided to deploy over 100,000 5G cells for business & competitive reasons. The main impact will be to save the telcos (and cablecos) a lot of money. 
     There's no evidence I can find that most of that savin will be spent on the network. Government "incentives" are giveaways to shareholders unless most of the money goes to investment. Otherwise, it to shareholders. In this case, the savings on the cells already planned or predictable will almost certainly dwarf the amount of additional investment. 
     Any reporter - or Congressman - hearing about "incentives" needs to ask the question, "How much more will be built?" Almost always, the answer will be disappointing if you get one. 
     The New York Chapter of the Internet Society resolved April 2 "Resolved: That ISOC-NY supports the letter of New York City Chief Technology Officer re: FCC Broadband. We urge the FCC not to move forward without majority support from non-corporate members. We also recognize that getting rid of unnecessary regulations on small cells is a good idea."
     I made the motion. 


  • Larry Press pointed out investment gains are not the only argument for killing neutrality. Larry is completely correct there are legitimate arguments against neutrality. Most of them begin with, "Getting government involved often creates problems." We all can accept that. We now have capex projections from the major carriers. They correspond closely to what they planned before repeal.
  • The brilliant Andrew Odlyzko is surprised by my thought "Even if I can't think what I'd do with 10 gig." He believes lower latency is crucial. I need to find numbers before responding. More to come.


  • Good news from Adtran. Tom Stanton reported strong demand for VDSL in Germany. AT&T is concluding the qualifying work for in district and will soon deploy it across the country. Australia’s is on track for later this year. Adtran’s NG-PON2 is in final testing at Verizon. Unfortunately, Century is still on hold, which resulted in a second straight dismal quarter.

press including tweets

  • Geoff Huston: the currently filed satellite plans encompass the launch of more than 13,000 spacecraft in the near future. Most are low or middle orbit with low latency. That may or may not be enough to replace almost all universal service programs.
  • Matt Blaze: Seeing a lot of glib "just don't be on Facebook" comments. I'm not on FB myself, and there's a significant cost to it. "Don't be on FB" is unhelpful, unrealistic advice for a large number of people. It's like saying "just don't have a credit card, or a car"
  • Dean Bubley: Definitely seems to be a bit of turf war coming in network automation and "zero-touch" between @ETSI_STANDARDS @tmforumorg @MEF_Forum & assorted other telecom/NFV groups Far from clear what the "ends" are in supposed "end-to-end" orchestration..


April 10

A special report: Who Are The Three Billion? The Color Of The Net Has Changed
1/3rd wireless only, 2/3rds in Global South, 75% not native English speakers.

Reply "subscribe" for a free sub, "un" to be dropped. Forward this to a friend. 

Xavier Niel became a billionaire with a 30 euro triple play and 100 gigabyte 4G for 20 euro. In an historic move, he is selling 10 gig fiber triple play for 50 Swiss francs ($53/month.) That’s less than Verizon charges for 50 meg naked and significantly less than almost all the 1 gig offerings. 
    His profit margin will be better than on 90% slower fiber at 30 euro. The large volume prices of equipment ($4 & up) are now low enough that most new builds should go 10 gig. 
    Even if I can’t think what I’d do with 10 gig.
More than 1B people are wireless only to the net, including hundreds of millions in India, Africa, and Indonesia. Assuming two people per wired connection, the 950M wired connections serve 1.9B. 
    65% of the total ~3B people are in the global South. Only 25% are native English speakers. About 1/3rd of the Internet is Chinese. Outfits like ICANN & the Internet Society are living ten years in the past. much more at end and follow the links.
Mark Zuckerberg wants to be a mensch, a mutual friend tells me. He truly believes in connecting the world, which he doesn’t understand.Beyond the DC circus, Indonesia is threatening 12 years in prison (won’t happen,) top Brit cop Shaun Sawyer believes “victims of slavery and human trafficking could sue the internet giants” (London Times,) and Julia Angwin has devastating reporting that FB would target ads to “Jew haters” while excluding blacks and Latinos
    Finally he agreed to pay taxes and fees in Korea, a story I broke in the West. (below)FB & Google do not pay taxes nor hire locally in innumerable countries around the world. The drain on poor countries is now in the $billions and creates hatred for America.
Donald Trump may be doing more for advanced Chinese industrialization than anyone since Deng Xiaoping. Cutting off China forces them to produce their own. They are putting $100B into memory chips alone. They just redoubled effort in 5G. They’ve made sure nothing goes into wireless standards without their approval. 
    Donald apparently believes China needs the more advanced U.S. technology. B_______. Huawei and others are as good as and often better than anyone in the West. Ericsson & Nokia were at least a year behind two Chinese on Massive MIMO. Baidu, Alibaba, and Tencent hold their own with Google, Facebook, and eBay. 
    Huawei is winning more patents than almost anything else and just won a major suit against Samsung. I've met some of the leaders at Huawei. They are more public-spirited most businessmen in the West, whatever the government.
The Brooklyn 5G Summit April 25-26 has two dozen of the very best engineers in the world. After two days, you'll know more than 90% of the pundits. Say hello to the round fellow with a beard and Jennie Bourne, The Woman With A Movie Camera. Or watch the live IEEE stream 

5G is becoming the new buzzword for policymakers, businesses and journalists, the meaningless technobabble-du-jour. Dean Bubley

10 Gig, Triple Play, $53/month. Xavier Niel, Switzerland
Xavier Niel's Salt is offering 10 gigabits to 1.3M Swiss homes for 50 Swiss francs, less than Verizon wants for 50 megabits. Salt includes an Apple 4K TV in lieu of a set-top, 300 TV channels, unlimited calls and the 4K Apple TV in lieu of a set-top box. 
    Xavi changed everything with his 30 euro triple play in France and became a multi-billionaire. Electric companies and munis have the fiber in place, with more than half the country soon to be covered.
    10 gig probably costs Salt $4-7/month more than GPON over several years, covering additional bandwidth and the cost of the equipment. The numbers suggest any large company with competition should be going for 10 gig rather than 1 gig GPON.
In Japan, ISPs from KDDI and Sony have brought 10 gig home down to less than $100. I thought that was a big story until the news from Bern. Adelaide is building 10 gig to 1,000 buildings in the central business district and major commercial strips. They expect to finish in two years. I'd recommend it to other cities with commercial buildings without fiber. Unlike Smart City manifestos, the high speeds for businesses are more than pr. More
The cost to build to 10 gig is low enough that it pays off even at a low take rate, if you have the fiber. 10 gig is about enough to serve every apartment in Jennie's 20 story building. It's enough for 30 4K camera raw outputs or about 600 4K video streams. More, including a look at the costs 

*** ASSIA is proud to announce that Telefonica has deployed CloudCheck® software solution across Telefonica’s Latin American Networks. The CloudCheck solution is a highly scalable, big-data-analytics, and machine-learning platform that that proactively detects and resolves subscriber Wi-Fi issues while automatically optimizing Wi-Fi performance for their subscribers. (Ad)

AT&T Shocker: 5G mmWave Latency 9-12 Milliseconds, Not 1-5 ms
1 ms will be in the lab for years.
    From China to AT&T to the EU, regulators are being promised ultra-fast, 1 ms latency,5G. They are being led by the nose to offer giveaways and subsidies. 1 ms. is certainly possible, as is 2 ms. LTE. But the networks being built are 5-15 ms., sometimes more. 
    In Kalamazoo, 1 gigabit went 300 meters. In Waco, the latency was 9-12 ms. over 150 meters. They observed no impacts due to rain, snow or other weather events. They tested 400 MHz signals.

*** Sundeep Rangan Understanding Ultra Low-Latency in 5G Cellular April 19, 2018 2:00 PM ET IEEE ComSoc Register for free (psa) Rangan of NYU Wireless is a world-class engineer. 1 ms. is possible - one day.

Editorial: Ajit Pai Needs To Stop T-Mobile-Sprint Before It Starts
A few simple comments like, "It's not clear the merger would be approved" would probably scare off Wall Street. The evidence from Europe is strong going from 4 to 3 wireless carriers usually raises prices 10-15% and cuts investment.
Pai said making broadband affordable is one of his two highest priorities. Since he took office, the price of broadband has gone up. If he doesn't stop the merger, it's 99% sure his work at the FCC will result in higher prices.

Historic: Facebook Will Pay Taxes and Internet Tolls in Korea
Kevin Martin of Facebook met the Minister in Korea and agreed "to report revenue and pay taxes" in Korea. In addition, they will probably pay "network fees" to the three big Korean telcos. Korea has allowed their telcos to erect "toll barriers" on the net, what Net Neutrality was supposed to prevent in the U.S.
    Tax avoidance is a huge issue around the world, especially as the Internet becomes a larger part of every economy. Facebook and Google are collecting tens of billions. often from poor countries, but neither paying taxes nor obeying local laws. Europe is the most visible resister, but the feelings are strong in Africa and Asia. "You Americans are so greedy," an African said to me at the WCIT in 2012.
    The web made tax evasion much easier, and the American giants own the web. Depending on whose numbers you use, Facebook & Google have between 60% and 90% of the ad revenue of the web outside of China. Google and Facebook are currently under tax investigation from Ottawa to Slovakia to Tel Aviv to Singapore. Italy and France have claimed $billions, the EU fines are about $3B 
   No one reading this needs me to rehash the arguments over neutrality. The Koreans are probably making a mistake, but it's certainly their right to decide policy for Korea. 
     Google is refusing to negotiate. Much more

*** Sckipio GFAST. Ultrafast broadband without the sawdust. The way to get to ultrafast broadband without all the disruption. Delivering up to 1.5 Gbps today. (Ad) With cable going to a gig, anything less than is very risky. db

Gigabit Broadband Downstream Available to > 50M U.S. Homes
80% coverage at Comcast leads the way. Over 100M in two years. Comcast, the largest with 54M homes passed, is headed close to 100% by yearend. The second largest, Charter with 49M, is beginning to turn on the gigabit and will be deploying heavily in 2018 and 2019. #3, Cox, is planning complete coverage soon.
    In a few years, about 40M will be able to get 1 gigabit symmetric from the telcos. Verizon has essentially frozen at around 15M, which AT&T intends to match by 2020. Millions more are fibered at smaller companies.
     The upstream for most cable will be 5-20 megabits, with a few able to receive 50 megabits. Higher cable upstreams will mostly wait until 2020 or later. John Chapman of Cisco tells me full duplex DOCSIS will yield hundreds of megabits upstream initially and gigabits eventually. He expects first units in 2019 and volume in 2020.

*** Brooklyn 5G Summit is an event not to miss. April 25-26 Sponsored by NYU Wirelesss and Nokia. IEEE Livestream. (psa) More than a dozen of the world's best wireless engineers are coming. See you there. Dave

China Unicom Down`~1M in December, Coming Back in February
Update: 500K adds in Jan 2018. In December 2017, China Unicom fell from 77,503M to 76,539M broadband customers. That's still the third highest in the world, three times as much as Comcast. Competition is a natural explanation. China Mobile has been adding ~3M/month for the last six months and 35M on the year. China Telecom added 660K in December and 10.4M for the year. China Telecom now has ~150M subs and China Mobile 113M. Nearly all (>90%) at CT & CM are on fiber all the way home. 
    China has > 800M LTE subscriptions, with many of the 200M 3G customers soon to migrate. LTE speeds are often 50-150 megabits. It's inevitable that some people will go wireless only; 5M have chosen wireless for convenience in Japan even though fiber was available. But China Mobile and China Telecom continue to grow rapidly on wires.

U.S. Q4: Cable + 717K, Telco -193K; U.S. 2017 Cable +2,720K Telco -626K
2,095K adds to 95,056K. Over 100,000,000 including smaller companies. No surprises here. DSL - especially lines not upgraded - continues to be clobbered. Most U.S. cable is 100 megabits or more. Almost half the cable homes are offered a gigabit. In my Manhattan building, Verizon doesn't offer more than 3 megabits.
   AT&T is apparently the anomaly here, adding 114K while the other telcos fell. However, without the 3M fiber homes and the joint marketing with DirecTV, they too would be negative. 
   Century, Frontier, and Windstream lost between 5% & 10% of their broadband lines. They've starved the network in order to keep the dividend high and eventually it caught up to them. Leaders at all three have been promising investors they would turn around the 600,000 lines lost in 2017. Given the inferior networks, I believe they are lying to themselves. Details

"Don't Buy From Huawei," The Senator Said
Fear of China began before Trump. The AT&T phone cancellation is just the latest unproven security accusation. Nearly a decade ago, I brought Huawei in to bid in a modest mobile network I was consulting on. We needed a better price for microwave backhaul gear; to get into the account Huawei came in about 30% lower for comparable equipment. Then the call came in to the company President from a Senator's aide. "Don't buy from Huawei," he said, and reminded the CEO how dependent the company was on government funding. There was no choice.
    There is absolutely no public evidence that Huawei is a security risk. Their connection to the Chinese government is weaker than the connection between Lucent and the U.S. De facto, Ericsson & Nokia actively cooperate with the U.S. gov to assist surveillance. Although nothing is public, it's possible the "security issue" with Huawei is not that they leak to the Chinese but that they refuse to work with the U.S. NSA.

Possible correction
DT probably hasn’t cited $400B as the cost to cover Germany with 5G. Apparently, I misread a DT estimate for Europe. Their figures are still much too high.


  • Benoit Felten, looking forward to April’s Broadband Asia, pointed me to his article about the emerging Asian telecom groups. I hadn’t realized that Malaysia’s Axiata has ~350M subscribers across 11 countries. Singapore’s MyRepublic is working with shared access networks in Australia & New Zealand. With an investment from France’s Xavier Niel, they could become very interesting. Viettel now has 90M customers in Laos, Cambodia, Haiti, Mozambique, Cameroon, Peru, Burundi, and Tanzania. Benoit has been a crucial analyst as fiber has expanded internationally.
  • Gabriel Brown of Heavy Reading noted, “Some of your 6G stuff (e.g. no-cell RAN, above 40GHz, shared spectrum, etc.) was supposed to be 5G. Whether that happens in practice, I’ll guess we’ll see in time. How and when the 5G to 6G threshold is crossed will be interesting.” I don’t have clear answers either. 5G in 3GPP goes to 54 GHz, but interest above that is increasing. Shared Wi-Fi spectrum is common, but the politics of sharing current telco spectrum is tough.
  • Grahame Lynch wrote about the Huawei ban. “In Australia, we banned Huawei from the NBN and bought from Alcatel instead. Their gear is made in Shanghai in a JV with a state owned company. You couldn't make it up.” I understand Nokia/Alcatel/Shanghai Bell has a very strong Communist Party cell. 

Press & analysts

  • Craig Moffett writes, “Sprint doesn’t appear to have a path to sustainability on its own.  With a significant step-up in their network capex (of an incremental $1.5B to $2.5B YoY) beginning in F2018, it is hard to see Sprint coming close to free cash flow breakeven – much less cash flow positive – in the foreseeable future. ...  there is perhaps an 80% probability of an attempt.  With the probability of regulatory approval at 50%, the odds of deal consummation are therefore something like 40%.”


  • Over 3,100 Googlers signed on to a statement, “Dear Sundar, We believe that Google should not be in the business of war. Therefore we ask that Project Maven be cancelled, and that Google draft, publicize and enforce a clear policy stating that neither Google nor its contractors will ever build warfare technology.”

Who Are The Three Billion? The Color Of The Net Has Changed
About 65% of 1.5B Internet "connections" (~ 3B active users) are in the Global South, led by the BRICS. About a third are "wireless only," as wireless speeds often are > 10 meg, caps are going higher, and countries like Indonesia and India are over 90% 4G. 
    Three quarters are not native English speakers. More Indians (250M) are on Facebook than Americans. Vietnam has the fastest growth. Indonesia has more than 100M "wireless-only." China has three times as many active users as the United States. 
    The gap between North and South is widening by over 60M in 2018. Six large developing countries grew at 5% or more in the last year, compared to only one in the developed world.  Most developed countries are between 75% and 95% connected, with little room for growth.
    In the early years of broadband, America & Europe strongly dominated. Then Korea and Japan boomed, passing the U.S. by some measures. By 2013, North and South were about even. This year or next, it will be 2-1. A four part article with the data.

The wireless data is very weak, so I'm using low estimates. 

What do I mean saying Global North & South? Since at least 2012, there has been a clear split. The split is not precisely geographical. Australia usually votes with the North and Russia with the South. The BRICS have been backed by almost all of Africa, Latin America, the Arab World and Asia. Sometimes It's described as "The Developing World" vs "The Developed World." It's taboo to say "Rich Countries" vs "Poor Countries."
     The essence of the debate is whether the decisions should be made by existing institutions like ICANN, 3GPP, and the giant multinationals ("the market") or by institutions more representative of the world Internet today.

Reply "subscribe" to be added, "un" to be dropped
Volume 18, #14  April 10, 2018



March 21


Reply "subscribe" for a free sub, "un" to be dropped. Forward this to a friend. 

Minister Wei insists China will be first in 6G. What's 6G? I've started the 6G Report, with a working definition of 6G as important advances in wireless that weren't ready in time for 5G. BT  Qualcomm
Meanwhile, 5G hype has been taken over by Alice's Red Queen. As much as 90% of the "5G" in the next few years will be sub 6 GHz, with little more performance than LTE and hence no major economic impact. Politicians worldwide proclaim 5G will change the world. Not! 
    John Stankey, AT&T's #2, instead calls 5G "The coming evolution."  Not revolution.
    Update: For example, Meredith Baker, who speaks for Verizon and AT&T, just implied that 5G will go 200 gigabits, doctors will perform surgery from the beach, and that 3.5 GHz is not 5G.   
China Telecom promises 2M 5G base stations under 6 GHz A bigger telecom story is hard to imagine. Very little will be the original 5G millimeter wave that is 20X faster with latency of 1-10 milliseconds. It will be a decade before the wondrous 1 ms latency is common. Probably more. LTE can be almost as fast. The latest from Ericsson: "LTE will have the capability to reach a latency down to 2ms on the air interface." Huawei and Nokia have demos. 
Donald Trump is very, very afraid of Huawei. Protecting Qualcomm, a great company, from decimation is right. Trump’s motivation is not. Peter Navarro, WH point man, wrote a book, Death By China, with a knife tearing the American flag. Trump believes it. 

2M 3.5 GHz Cells for China Telecom
Chengliang Zhang's keynote at OFC revealed that China Telecom will build ~2M base stations for a national network at 3.5 GHz. That’s ~twice as many as today. Most will be built between 2020 and 2025. 
     Gigantic China Telecom and most of the Europeans are choosing "mid-band" including 3.5 GHz, ~90% of the announced “5G” is little faster than LTE.  CT's new backbone will be optical, fast, and almost failsafe. They are using 25G Ethernet for the edge rather than the NG-PON2 at Verizon. His goal is to have 90% of China within 30 ms of a CT server. The telco NGMN CTO's have a 20 year roadmap with the same goal: Everything through the telcos. China does not impose Net Neutrality and the system is designed to place almost everything on local telco servers. 
      This is the wet dream of most telcos around the world but so far never achieved.

*** ASSIA's CloudCheck® Wi-Fi Management Platform Now Supports Over 100 Gateway/Router Models. CloudCheck® API Enables Low Touch Agent Integration onto CPEs (ad)

LTE Latency Today 9 ms. Down to 2 ms ~2019
1-2 ms is on the 5G roadmap but years away almost everywhere.  All announced telco 5G is at 5-15 ms. The new LTE gear is just about as fast.
     These figures are from Ericsson, which has been working on 5G since 2011. All Ericsson radios since 2015 have been able to switch from 4G LTE to 5G with software. Nokia probably is the same or worse but I'm still waiting for the data. Depending on the upgrade cycle, LTE in some places will have lower latency than 5G at 3.5 GHz.
     "From mid-2018 Ericsson will have support for latency (Round Trip Time, RTT) down to 9ms on the LTE air interface (Based on the 3GPP Rel-14).... LTE will have the capability to reach a latency down to 2ms on the air interface." A shorter TTI (Transmission Time Interval) is key. 
     90% of "5G" before next decade looks to be < 6 GHz, with speeds almost identical to LTE.  5G NR couldn't have added much performance. 4G LTE efficiency was already close to Shannon's limit.

*** Sckipio's chips deliver a true gigabit. Twice the speed, four times the true gigabit ports, totally dynamic. The leader. (ad) 

What is 6G Wireless? (Part 1)
"6G" wireless is technology that is beyond "5G." Technologies moving from the lab to deployment should improve capacity by 5-15X over "5G." Three technologies will almost certainly be included, with more to come:
   • "Cell-free" or "Distributed" Massive MIMO
   • Dynamic use of different frequencies, especially unlicensed
   • Frequencies up to 120 GHz. Much more
More on 6G: BT's McRae: "Hopefully we won't need 6G." ; Qualcomm: Sharing Licensed Spectrum Would Work in 6G

AT&T's "Riverboat Gambler" Whitacre Likely Trump's Man in Mexico
“He was a riverboat gambler,” his colleague and successor Stephenson said. “It was anything but a risk-averse culture.” I like Big Ed, but whoever approved his security clearance needs to find a different job.
     He committed obvious perjury at the Senate. He refused to answer question after question. Senator Spector said, "That answer is contemptuous of this Committee," Whitacre just smiled. Everyone in the room knew AT&T conducted massive wiretapping without legal authority.
    Unlike Trump, Whitacre actually loves Mexico. Like Trump, he says whatever he feels like; his President, Bill Daley, would answer questions, "That's just Ed being Ed." Whitacre terrorized his staff; a very senior technologist told me it might be "career threatening" if he was seen talking to a reporter.
     He believes he is always right - even when he isn't. He expects everyone to take his orders; when he was CEO, the company was run top down like the army. He expected military-style secrecy. 3M Homes Passed, 29K Subscribers (Point-Topic)
Second generation chips set to speed deployment. Point-Topic, a very respected broadband data source, now has figures for They also have produced a clickable map. I believe they have all the large and medium-sized telcos with announced deployments. They've identified 93,000 homes passed at British Telecom, a disappointing figure if they are to reach 10M by 2020.
    Many in the industry aren't making information public so there is a lot of uncertainty. I know there are numerous smaller companies active; Calix has announced 100 active customers, some only a single building but some good-sized.

*** CommsDay announces CommsDay USA, the daily PDF digest for busy telecom executives. (ad) Good reporting is always welcome. Petroc & Grahame are among the very best. Dave

Colorado's Jeff Storey Up, Louisiana's Glen Post Out
Takeovers kill jobs, often affecting many of the best workers. If AT&T takes over Time Warner, New York will lose thousands of jobs. Senator Schumer could have resisted the deal but did nothing. Century, Louisiana's largest company, was likely to lose jobs to Level 3 in Colorado after the merger. Louisiana Senators John Kennedy and Bill Cassidy didn't even try to stop the deal. Century, Louisiana's largest company and America's third largest telco, will now be managed by Jeff Storey in Colorado.
     Century virtually stopped the badly needed upgrade of their networks after the merger went through. The unconfirmed rumor is that going forward, Century will move investment to the Level 3 corporate data network. Many parts of Century's network will be milked for whatever the company can collect before demanding a public bailout.
     Century stock has lost one-third of its value over the last nine months, so Storey probably had little trouble convincing the board to make the move. In a parting shot, the official photos have Storey looking sleek and Post badly in need of a shave.
      Century has a goal of 40 megabits to most by 2020. That's probably "up to 40 megabits" and actually less. Their fiber and plans, modest already, have been cut. They simply aren't competitive with cable, which by 2020 will have a gigabit to 90%.
       "We saw a net loss of about 90,000 broadband subscribers this quarter, which were made up of losses of about 140,000 at speeds under 20 megs and gains of approximately 50,000 in higher-speed offerings." (Seeking Alpha transcript)
       Frontier and Windstream, comparable companies, are down 80%+ to a price that only makes sense if bankruptcy is close.

From The Source: Balázs on What's in 3GPP 5G NR
Balázs Bertényi of Nokia is the 3GPP RAN Chairman, who recently oversaw some remarkably intense work to finalize "5G NR." Guy Daniels of TelecomTV got clear answers to questions every pr person has been ducking: What is it exactly? How does it differ from LTE radio?
    Bertényi's "two main elements" of NR apply to high frequencies only. 5G NR operates in higher bands, such as 28 GHz millimeter wave. There's a great deal of unused spectrum up there. Also More bandwidth in wider channels, up to 400 MHz. Few today have more than LTE's 100 MHz limit. But the U.S. 28 GHz band has ~2,000 MHz. I recommend you listen to the video. It's less than three minutes.
     Bertényi doesn't forget, "We also have Massive MIMO." He projects the improved version of MM in Release 15 has 15-20% more throughput in the lower bands. I'm researching whether the improved MM is because of the "New Radio" or just is another improvement in Release 15.
     With China Telecom now committing to < 6 GHz, the strong majority of 5G in Europe and Asia will not have performance much better than LTE for years.

*** Brooklyn 5G Summit is an event not to miss. April 25-26 Sponsored by NYU Wirelesss and Nokia. IEEE Livestream. (psa) More than a dozen of the world's best wireless engineers are coming. See you there. Dave

Trump And Putin Want To Control Internet Governance
Two of the most powerful men in the world want ICANN to disappear. ICANN and friends are in San Juan partying like it's 1999, thinking Putin is impotent and Trump's NTIA chief, David Redl, should be ignored when he promises to move on retaking control of ICANN.
    Putin personally ordered his government to "Develop an independent internet infrastructure for BRICS nations." On the independence of ICANN, Trump's Commerce Secretary, Wilbur Ross, is looking for "any possible [mechanism] for reversing it.
    David Redl at NTIA is tasked with delivering on Trump's campaign rhetoric, which opposed giving control of "the internet to foreigners." His nomination was held up for months by Senator Cruz, until he promised, "to investigate options for unwinding the transition,"

WTF! China Hawks Want 5G Gov Net. Ain't Gonna Happen
White House nuclear expert Gen. Robert S. Spalding II and Steve Bannon's guy Peter Navarro shocked D.C. with a serious plan to build a national 5G network in 400 MHz of 3.5 GHz spectrum.
     Even "progressive" Democrats like Jessica Rosenworcel and Colin Crowell immediately asserted the government building a 5G would be a mistake. Conservatives like Pai were vehement in their objections. A government-built 5G network is politically impossible even if the plan were a good one.
     WTF is going on?  Spalding and Navarro are hardline Chinaphobes. The illustration of the knife through the United States is from Navarro's Death By China. I suspect these guys think Nixon made a mistake recognizing China in 1973.

Worth knowing:

  • Top 15 telecom operators in the world have experienced a decline in capital expenditures, C114
  • “by mid-2018 Ericsson will have support for latency (Round Trip Time, RTT) down to 9ms.” 9ms is good, but isn’t 1ms  
  • Cable can tweak DOCSIS to bring backhaul under 5 ms
  • DSL & GPON sales were down in 2017, Alan Tamboli reports But Julie Kuntsler sees good growth the next 5 years. Broadcom sees "In Q2, very strong revenue recovery in our broadband carrier access, which basically reflects DSL, digital subscriber line, PON, and attach enterprise and carrier Wi-Fi. ...  very, very strong. ...But if you take it on an annualized basis, I would say, it’s practically flat."


  • Simon Stanley at Heavy Reading predicts “ is coming of age with the latest Amendment 3 chipsets.” I can confirm the vendors are confident that the results coming from the field will erase doubt.
  • “Capex as a percentage of revenue will drop from a current level of 15% in the cable industry to around 10% over that five-year span. Meanwhile, capex per home passed will fall from $140 to $120.” Jonathan Chaplin’s remarkable forecast, reported by Daniel Frankel This would be the most significant cut in capex in the last decade. It’s not impossible; half of cable capex today is set tops. Network DVRs and intelligence drastically cuts the cost of the box at home. The great unknown is whether cable builds their own 5G networks. The engineers are ready and excited. The business model almost certainly works: they have networks and customers that will bring the cost down. But one of the telcos might preempt the build by making an offer too good to refuse. Almost all telcos have more capacity than they can sell, likely for the next decade. The technology is improving that fast.
  • Steve Glapa of Tarana Wireless is enthusiastic about the major telcos looking closely at their wireless system for 3.5 and 5.8. He’s confident of remarkable performance from interference cancellation using eight antennas at the consumer site. That’s related to 6G distributed MIMO. I’ll be watching closely as they release verifiable data.
  • From MWC, I read more than 300 5G tweets and dozens of articles. Many picked up the key news, the important two gigabit chips. Beyond that, I think I only read two that weren’t straight from the press release or speech, unimportant, or previously known. Disappointing. At an event with so many senior engineers, you can almost always find stories simply by asking a few questions about important stuff that isn’t being discussed.

Reply "subscribe" to be added, "un" to be dropped
Volume 18, #13  March 20, 2018

Feb 27 MWC Wireless special 

Höttges of DT, Mittal of Bharti, Patterson of BT, Qualcomm’s Mollenkopf, Ajit Pai of the U.S. FCC, as well as Ansip and Viola of the EU, and many more offered preposterous claims. Very few noticed the (metaphorical) King's New Clothes.

Most 5G is no faster than 4G was the big story at the show. That is not a typo. Qualcomm and Huawei brought remarkable new chips that probably cost $B to develop. 

Huawei's chip for 5G at 3.5 GHz delivers 2 gig peak and 200-600 typical in the real world. 
Qualcomm's chip for 4G LTE delivers 2 gig peak and 200-600 typical in the real world.

Explanation: 28 GHz is the 5G with benefits. 3.5 GHz (Majority in Europe) is 4G LTE with a software tweak (NR) and 100's of press releases. It's 70% to 90% slower. The benefits the telcos claim and even honorable politicians believe are mostly hogwash, 

More not to believe: Patterson of BT has said it’s financially impossible to bring fiber to most English homes but Telefonica and Orange/FT have done it. TEF and ORAN stocks have done 40% better than BT the last few year; DT is quoting 5G costs ten times what Verizon estimates. They want a massive subsidy; Viola thinks 100% "5G" is "crucial" for self-driving cars, but millions of fully autonomous cars are about to hit the roads using radar & lidar.  More at the end or

First. Independent tests of VZ mmWave in Houston by Thelander found mmWave is going twice as far as expected, even with obstructions (NLOS.)Then, new 2 gig in depth.

2,000 Feet Field Test of Verizon 5G, NLOS Often Close to a Gig
Verizon 5G reach is perhaps twice as far as expected, including with obstacles (NLOS.)

That's the independent conclusion of Michael Thelander and team testing two live Verizon cells in Houston. 
"Who would have thought a millimeter wave signal in an area 100% blocked from the serving cell tower by the surroundings would still be capable of supporting good data speeds?" they exclaim. 
    "Millimeter wave signals are far more resilient than we expected, even at distances exceeding several thousand feet. Tree foliage, passing school buses, buildings, parked cars, balding heads, and glass impacted the received signal, but the resultant signals were still capable of delivering meaningful data rates – thanks in part to the 400 MHz radio channel.” 
     The results were generally good up to 1,000 meters and behind buildings. Dozens of analyses of 5G costs assumed reach of 200-300 meters. They probably are wrong. more

*** ASSIA's CloudCheck® Wi-Fi Management Platform Now Supports Over 100 Gateway/Router Models

CloudCheck® API Enables Low Touch Agent Integration onto CPEs (ad)

2 Gig Chips Revealed: 4G LTE, 5G Essentially the Same
Typical speeds: 100-600 megabits down, much less up. 

     The new "2 gig peak" chips from Huawei and Qualcomm deliver about the same speed with a similar number of antennas and the same spectrum. You can't get much more speed without breaking the Laws of Physics, whether you call it 4G or 5G.
    The real 5G uses perhaps ten times as much spectrum for practical speeds over a gigabit. You can only find that much spectrum free at millimeter wave ranges, mostly 26-30 GHz. I and most others consider Massive MIMO, 64-256 antennas, also 5G.
     Early in 2017, the marketing people started to call 3.5 GHz "5G" even though the speed was almost the same. The result is what Dan Jones calls Faux 5G and I call Fake 5G.
     The software tweak (NR New Radio) did little for the speed although NR offers other features. The marketing people love the name 5G and stuck it on everything. 
    Since the 2009 first LTE networks forward, LTE in ideal conditions delivers 90+% of the Shannon capacity per hertz of spectrum per antenna. The first generation of LTE used 20 MHz and 1 antenna. Moore's Law and improved components allowed increasing the spectrum used and adding antennas.The new chips use 5-8 times as much spectrum and 4 antennas. much more

*** Sckipio's chips deliver a true gigabit. Twice the speed, four times the true gigabit ports, totally dynamic. The leader. (ad) 

Qualcomm 2 Gig 4G LTE
The Snapdragon X24 LTE modem has a theoretical peak of 2,000 MHz using ?140 MHz of spectrum and 4x4 MIMO. In the real world, expect 100-600 megabits typical. 

     This is a heckuva chip, one of the first 7 nanometer chips produced. While none of us require more than 100 megabits 99% of the time, the total capacity is also important. The speed will be divided among the many users of the cell.
     Real world speeds vary from lab speeds because cells will have more than one user, greater distances, varied topology, and other interference. There will be times when you are the only user and in the exact right spot and can get close to the peak speed. It's likely that users at the edge of the cell will see the greatest improvement.
     140 MHz very rare outside of the laboratory. Verizon and AT&T only have about 100 MHz, total. They expect to take over the Wi-Fi spectrum. There's also about 400 MHz around 3.5 GHz in the pipeline. I don't believe the RF front ends needed exist yet and very few phones are capable of using all this capacity. 

2 Gig 5G from Huawei
The $B Balong 5G01 is a remarkable achievement, with versions for millimeter wave & low/mid band 
     They are mainstream, 3GPP standard chips very similar to the advanced chips of leaders Qualcomm, Intel, and Apple. The processor cores are ARM. They showed off working routers for both < 6 GHz and mmWave. These are first generation chips; second generation, due later in 2018, will be smaller. They expect routers for sale this year and phones in 2019.
    Most tech specs are not yet released, including how much spectrum is needed for 2 gigabits; power consumption; heat generation; and practical latency. 
     I estimate the full development of this chip at $1B, based on how many engineers are needed for 5G chips and the costs of the necessary 7 nm chip production. TSMC is probably producing this chip (and Apple's.) Intel and Samsung may be the only other fabs in the world capable of the level of complexity. Both have their own 5G chips on the way. 7 nanometers is getting close to atomic sizes.
     Contrary to rumor, Moore's Law isn't dead yet. 5 nm and probably 3 nm chips are coming in the next few years but they won't be cheap. Both TSMC and Samsung are spending $20B on single plants.

*** CommsDay announces CommsDay USA, the daily PDF digest for busy telecom executives. (ad) Good reporting is always welcome. Petroc & Grahame are among the very best. Dave

Confirmation: 28 GHz 5G 1.4 Gbps Median: 3.5 GHz 5G Massive MIMO 490 Gbps
5G at 28 GHz (millimeter wave) is a breakthrough, > 5X the speed, < 1/3rd the latency. 5G at 3.5 GHz is 70% to 90% slower, about LTE speed. 

     Qualcomm's simulations are respected and the most accurate predictions available. They are consistent with other data. I've come to similar conclusions from expectations of the technology. One possible update to come: The results from Verizon's tests at 28 GHz are very good so far, and may raise performance estimates for millimeter wave. 
     Topologies vary. There's no way to predict perfectly the number of users, distances, window and tree interference, window glass, trucks passing by and other crucial factors. Simulating wireless networks is an art, not a science, and requires great skill.
    Good models are crucial for planning actual deployments. I've used the actual Alcatel and Ericsson LTE deployment models, which are very rich. Everyone in the business has learned models will never get everything right, but are an essential starting point. 
     The industry knows problems are possible from bad projections. Australia's NBN is several hundred million dollars behind due to errors in modeling their cable networks. Many customers had serious problems. They are rebuilding some sections and delaying others. 
      ADSL was designed in 1998 to deliver 1.5 megabits downstream 18,000 feet. Telcos committed $B's on that basis. The results were disastrous and most companies revised plans based on 12,000-15,000 feet. For a company like Baby Bell Ameritech, the overrun was probably in the hundreds of millions. The New York Times called it "DSL Hell" because of the customer problems. Top executives were fired. Cable pulled ahead in the U.S. Telcos spent $billions to catch cable but remained behind for the next decade.
     Five years from now, with real results, I'll discover major errors in the articles I'm writing about 5G.

Update/correction: Contrary to press reports.Comcast spokesman Harold Ford was not fired by Morgan Stanley because of improper advances, he and the company announced after lawyers got involved.


  • Zahid Ghadialy, a good reporter, tweets, “I have to be honest, haven't seen a WOW demo yet at #MWC18. While there are lots of interesting stuff, it's all the same, old and tired stuff.” I’d make an exception for the 2 gigabit LTE & 5G chips, partly because they provide a good way to show people that 5G at 3.5 GHz is no faster than what’s called 4G LTE.
  • Unless intelligence is moved to the edge (URLLC) 5G latency will be 5-10 ms, not the widely touted 1 ms. Mike Murphy of Nokia says, “Carriers globally are at different stages that range from not thinking about this yet to having all their decisions made and being on an execution path.” I’ve looked hard but haven’t found a single telco on an “execution path.” Everyone’s looking and some will trial URLLC but the massive cost is deterring them. Rick Merritt, one of the best, has an excellent article .
  • ZTE is issuing $2B of stock to finance their $6B of investment into 5G.
  • A single atom is this remarkable photo

Naked Emperors At MWC Barcelona
Big errors obvious from 3,000 miles away 

Höttges, the first MWC Emperor without Clothes, wants $10’s of billions of subsidies and other favors. To get them, he claims a cost ten times higher than Verizon.  VZ expects a U.S. 5G cover will cost $50-$80B or less. DT's $400B to cover Germany (pop 83M) (pop.326M). Another unlikely claim is that 65% of Germany will have 100 megabit fiber this year. "Up to speeds: 50% of the customers on a similar network (without 35b) in Australia don’t get 100 megabits and 20% don’t even reach 50 meg. note: Höttges told investors DT will have 14M lines of 250 meg DSL by the end of 2018, only a year or two behind schedule.) 
Mittal has built Bharti to 380MM subs. They gained share this year despite Jio’s $20B challenge. Remarkable achievements. But must the government subsidize a $9B man because his costs go up by 3% of the $25B market cap. Bharti is convincing analysts and reporters rich companies should be subsidized even if they don’t invest more.
Patterson of BT has said it’s financially impossible to bring fiber to most English homes.  I don't believe BT engineers are any less competent than those at Telefonica and Orange, which have already fibered most of France and Spain. It's clearly possible without major financial problems. TEF & ORAN stock prices have done 60% better the last two years. 
Pai of the U.S. (and Democrat Rosenworcel) think mmWave 5G will offer 1 ms latency while Verizon and AT&T are building for 5 ms to 10 ms. I don’t believe any telco in the world has announced 1 ms. networks. Moving the intelligence to the edge (URLLC) is brutally expensive. He insists killing Net Neutrality will raise investments, but the evidence is the effect, if any, is too small to notice. Most of the capex plans for 2018 have been announced and they are the same as they would be with NN.
     However, I deplore the very cheap attacks on Pai. He’s a hardworking, brilliant lawyer. He does what he does because of his politics, not because of corruption. He took a courageous stand on immigration, which could have cost his job.
Mollenkopf of Qualcomm was busy trying to get a higher price for the company’s soul. He claims 5G will add $11 trillion to the economy in 2035, but almost everything attributed to 5G in his IHS “study” can be achieved with 4G, including by Qualcomm's excellent new 4G chip. (I double checked.) Any competent economist would conclude a study like this for 2035 is a mistake. Even five-year projections would be tough. Qualcomm engineers are the best but his lobbyists are as trustworthy as Bill Clinton discussing Monica Lewinsky. The LAA chip demonstrations prove it does harm Wi-Fi despite multiple times Q said the opposite.
Ansip and Viola of the EU are smart, hard-working and I believe honorable. They are either blinded by ideology or just don't understand telecom. They tweeted 5G will offer 10 gig connections. 1 gig is the target. 10 gig is the total system, split between many users. They are badly mistaken that 5G requires 25-year spectrum licenses at a "bare minimum." Not one of the telco CEOs who make that claim has said they will build more. They demand incentives for someone else to build, not very likely. Incentives are a mistake without evidence of probable investment. I've looked at the data and it proves no such thing. 
    LOL at Viola's tweet that implies we won't have autonomous cars for a decade or more. Google has fully autonomous cars carrying passengers today in Phoenix and has ordered a fleet of thousands. "100% 5G coverage" is clearly not a "key condition." Europe will not have wide 5G at least until 2025. 2030 almost certainly will not be ~100%.
     Hundreds of thousands of fully autonomous cars will be on the roads soon. They depend on radar and lidar, not the telco. If you haven't realized just how soon these cars are coming, do watch the video of Google's John Krafcik Unbelievable demo at 6 minutes. (There's plenty still to get right, but they work.)

If I have any of these wrong, please do email me, This email address is being protected from spambots. You need JavaScript enabled to view it.. I make many mistakes, as the butler said to Humphrey Bogart. 

Reply "subscribe" to be added, "un" to be dropped
Volume 18, #12  Feb 27, 2018 300,000 Indian Villages Fibered; 325,000 More To Come
Verizon, NTT, AT&T: ?$200-$400 5G Costs Much Lower Than Expected
Important: 5G Handles 10x More Data Than 4G
India Passes U.S. in Smartphones
AT&T Probably Will Not Have a 5G Mobile Phone in 2018
Pai, Speaking of Immigration, Put His Career on the Line
35b: Ready or Not? Germans Say Not, Some Say Yes
Qualcomm's Not To Be Believed Comeback in

Reply "subscribe" to be added, "un" to be dropped

Manoj Sinha announced that 300,000 Indian villages have been connected by fiber and BharatNet is on target for 325,000 more early in 2019. This is by far the largest rural broadband project in history. Bids are out for 500,000 Wi-Fi hotspots and 130.000 post offices will connect. The telcos are offered low prices to bring in LTE and fiber home. Programs are underway for > 100,000 local business to offer connectivity.  

Ajit Pai risked his career by challenging Trump's immigration policy. "My love and reverence for this country comes from living in the house of Raj and Radha Pai. My parents know a little something about the American Dream. They came to this country 46 years ago with literally no assets other than $10, a transistor radio, and a desire to achieve that dream." I never imagined America would plan to deport four million people. 

  • Boldface correction: Adtran’s “Tier 1” customer who severely cut investment apparently was Century, not AT&T.
  • This issue is dedicated to Wa Lone and Kyaw Soe Oo, Reuters reporters in jail in Myanmar for reporting the truth.
  • A news story just making the rounds that Verizon is throttling  Netflix to 4 megabits is almost certainly a mistake. Netflix sends HD video at ~4 meg. An app thought Verizon had reduced the speed.
  • Beware Bitcoin true believers. Nobel Prize winners and anyone with common sense realizes the price is unsustainable but they can’t hear that.

Breaking: NY Times Sees U.S. Using Nuclear Weapons in Case of a Cyberattack
"A new U.S. strategy 'would permit the use of nuclear weapons to respond to a wide range of devastating but non-nuclear attacks on American infrastructure, including what current and former government officials described as the most crippling kind of cyberattacks. This would be 'first use."

300,000 Indian Villages Fibered; 325,000 More To Come
Takeaway: Sometimes, government gets it right. This is large enough to be world changing.
     After years of delay, >100M Indians are ready to be connected with fiber, Wi-Fi, and LTE. Phase 1 is finished of India's BharatNet, the most ambitious rural connectivity program in history. The expectation is all 250,000 regional councils and 625,000 villages will be reached by GPON fiber by the March 2019 completion target.
     Most villages will have local Wi-Fi, sometimes from the local Post Office and often through a program of local businesses. Backhaul for telcos is included; Jio will presumably use BharatNet backhaul as they extend their LTE network to 95% of Indians.
     The fiber and most of the other equipment is produced in India. State-owned BSNL is playing a major role in construction and procurement. The $15B project is primarily funded by the accumulated Universal Service Fund.
      China, Vietnam, and now India are showing what governments can do delivering broadband. (Sometimes) The majority of new Internet connections are now due to government entities.

*** Publisher's favor request. FNN has one of the most desirable audiences in telecom, loaded with senior decision makers as well as influential press and analysts. I'll design an ad package for you that delivers excellent value for money. No contracts or other hassles. Email This email address is being protected from spambots. You need JavaScript enabled to view it. or call me 347-603-6442 for some ideas you will like.

Verizon, NTT, AT&T: ?$200-$400 5G Costs Much Lower Than Expected
Hans Vestberg of Verizon is passing 30M homes without raising capex.
     NTT DOCOMO's very respected CTO Seizo Inoe in 2016 called high costs of 5G "a myth." He expected LTE to be cheaper than 3G and that much of the 5G would use existing towers and backhaul. Vestberg, former Ericsson CEO now running Verizon's network, sees costs of mmWave costing $200-400/home passed as they deploy to a quarter of the U.S., probably by 2020 or 2021. (Obviously, some other areas will be more expensive.) 5G is "Massively, massively cheaper than having a fiber all the way to the home to have sort of a beam in the air going to the home."
     Verizon has previously said their costs to pass a home with fiber were $700 in 2007 and went down from there, presumably to $400-600. I'm inferring that "massively, massively cheaper" would be $200-$400.
    All cost numbers like these are rough estimates and vary enormously based on density and physical circumstances. Verizon presumably went first to lower cost areas. The second quarter of the U.S. will be more expensive and the last quarter even more so. In particular, the last 2-8% of rural fiber costs into the thousands and the last ~1% is prohibitively expensive.
    The main reason the cost projections are coming down is that mmWave reach is proving much better than expected so fewer cells are needed. Vestberg is often getting a gigabit 300-700 meters. A major European government study was based on 200 meter reach. The high cost estimates for "5G" from outfits like Deutsche Telekom are pure b_____, designed to get government subsidies and concessions. 
    None of the above is certain until we have more data from the field.

Important: 5G Handles 10x More Data Than 4G
Takeaway: Wireless progress has and will keep up with demand without major changes in capex or spectrum. 3 out of 4 in the industry get that wrong.
    Vestberg also said: "2G to 3G, probably 10x better to handle data and 3G to 4G 10x at least. And minimum, I would say, 4G to 5G is the same." 
    "Will we be able to keep up with wireless growth?" I asked Hans Vestberg back when D.C. was screaming spectrum crisis. "Yes we will. I'm confident human ingenuity will deliver what we need," he replied. "That's always been true in wireless." 
     Five years later, wireless speeds worldwide are much higher; congestion in the developed world remains an exception. Telcos continue spending $billions on advertising to find enough customers to fill their nets.
     Without much more spectrum or many more cells, improved technology alone will cover likely demand until 2024 or 2026. (My calculation.) Putting to use currently fallow spectrum and modestly more density, wireless networks are good until 2030 or later.
     What Vestberg said confirms what I heard from Paulraj of Stanford in 2014. He believed MIMO alone will yield a 50-100x improvement. Vint Cerf, Henry Samueli of Broadcom, and Andrea Goldsmith generally agreed at other Marconi events.

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." (ad) Really does a gig and more, including upstream db

India Passes U.S. in Smartphones
~40M sold in Q3, increasing. 
     With > 400M Indians connected by 2G phones, India has become the second largest smartphone market and is gaining share against China. With monthly service prices often under $5 and many phones subsidized to $25-35, Deloitte estimates 350M more Indians will be connected by 2020. Reliance Jio added over 100M customers on 4G smartphones in the last year; the remaining survivors, especially Bharti & Vodafone, are scrambling to catch up.
     Google's new Oreo Go version of Android is bringing the cost of low-end 4G phones to about $30, with the first units expected from Indian manufacturer Micromax by the end of the month. The inexpensive phones will be a natural way to connect in the amazing 625,000 villages being fibered by BharatNet.
    The Chinese vendors, including Oppo & Vivo but not Huawei, have massively taken market share from Indian companies like Micromax and Karbonn the last few years. The Chinese have been moving production to India to beat 15% tariffs and find lower labor costs. Apple has talked a good game, but is holding out for enormous tariff and tax subsidies from India. More

*** Telefonica Deploys ASSIA’s GPON Expresse Solution
ClearView support for Wi-Fi and Expresse are now available for both DSL & GPON (ad) 

AT&T Probably Will Not Have a 5G Mobile Phone in 2018
Despite dozens of news stories, AT&T has not said they will launch a mobile 5G phone in 2018. 
     They carefully spoke about a "device," may be a tablet or laptop. They confirmed that when I asked.  A tablet or laptop has power, space, and heat dissipation capabilities that a contemporary mobile phone lacks. 
      It's not impossible to have a mobile mmWave phone chip this year. TSMC will be in "risk production" of 7 nm. Linley estimates the mmWave chips will require 10X the transistors of 4G, raising major questions of heat and power until 7 nm. Most telco engineers were amazed when Intel & Qualcomm moved delivery from 2020 to 2019. Delivering commercial chips in late 2018 would be remarkable. 
      T & VZ face the same issues, deploying the same equipment from the same suppliers. Verizon is doubtful. "It is a year before you have something at least and -- until you have it commercially as well. Maybe a little bit more than that when it comes to wireless."
     They both are racing for the pr; neither is likely to be much ahead. These are pr battles. There is a real battle coming to actually deploy across the country, if my sources are correct.

*** IoT Evolution Expo Orlando Jan 22-25
5 Collocated events, 80 Sessions, 250 Speakers, 2,000 Attendees & Unlimited opportunity (ad) Programmed by Carl Ford, one of the best

Pai, Speaking of Immigration, Put His Career on the Line
"My love and reverence for this country comes from living in the house of Raj and Radha Pai. My parents know a little something about the American Dream. They came to this country 46 years ago with literally no assets other than $10, a transistor radio, and a desire to achieve that dream." Pai is a brilliant lawyer steeped in politics. He surely realized his job could disappear if Trump noticed the speech Pai gave in honor of Ronald Reagan.
     Pai will go down in history as a man who killed Net Neutrality, or tried do. There's more to him than that, including the courage to stand up for his very strong beliefs. He's wrong in many of those beliefs: AT&T has made clear that killing neutrality will not lead to the increase in investment Pai expects. T-Mobile & Verizon will build in 3.5 Ghz spectrum without the concessions Pai is about to give them; much, probably most of the money he thinks will go to expanding broadband is wasted.
    But he's right that immigrants should not be scapegoats.

35b: Ready or Not? Germans Say Not, Some Say Yes
A reader wondered about my 35b story writing, "35b to my knowledge is already in market." I know Fastweb in Italy is already selling to consumers at "Up to 200 megabits" but without vectoring they will often be much lower. On the other hand, Deutsche Telecom, the largest expected customer, says it won't be ready until 2019. I believe what's going on is that Broadcom is selling 35b chips but the vectoring isn't ready for large deployments.
     The reader asked what were my sources for the story Deutsche Telecom: 35b Supervectoring Delayed to 2019? That's always a fair question. 
     I've been reporting 35b since 2015, but this story was inspired by an article by Achim Sawall. The official statement from DT did say 2018, so I checked with Sawall, a respected reporter. He confirmed to me that DT had updated the timing at a later press briefing.
    All of us are waiting for detailed results from the field.

Qualcomm's Not To Be Believed Comeback in
Qualcomm Ikanos division and product killed in 2016. This was widely reported and officially confirmed after I broke the story. A link from Google News took me to a "research report" from, with the four leading chipset vendors identified. They correctly identify three chipset vendors, Broadcom, Ikanos, and Metanoia. But I strongly doubt the "report" is worth its $2900 price given they claim Qualcomm makes chips. If anyone knows who is behind this fake news, please tell me.


  • Craig Moffett believes video price hikes “are like a car that is headed for a cliff, with a driver that believes the only solution is to step on the gas.”
  • Craig Matsumoto has joined 451 Research after 20 years as one of the best tech reporters. He writes, "Journalism was loads of fun, but after years of being told I ought to try being an analyst, I'm giving it a go. I'll be covering managed services but still hoping to find challenging technologies to write about -- stuff that's not only new but a little bit weird."
  • Great story: 17 year old fabbing 4004 chip in garage
  • Great headline: Nepal bans blind people and double amputees from climbing Everest
  • Sony and Softbank believe we will have robot companions soon; Pepper & Aibo still have a way to go to satisfy most of us but AI keeps getting better. A well-illustrated BI story about Gatebox AI gives a hint of what’s to come.

December 28

Reply "subscribe" to be added, "un" to be dropped

Akayed Ullah bombed the 42nd St subway tunnel I often use. Walking through two days later, I thought about the effect on our industry. I realized that ubiquitous surveillance cannot be stopped. Right or wrong, with bombs frequent a fearful majority will give Homeland Security almost anything they request.

Good news coming: Rwanda is at 92% LTE, going toward 100% in the next year. Both Myanmar and India are heading towards 95%. If they can do that, so can almost every other country in the world.
I believe Telecom Italia is under 90% but they refuse to give me a figure.

Apologies for the long Internet Society stuff at the end. With $30M/year from .org, ISOC should be the most important advocate in the world for Internet users. It isn't.

*** Publisher's favor request. FNN has one of the most desirable audiences in telecom, loaded with senior decision makers as well as influential press and analysts. I'll design an ad package for you that delivers excellent value for money. No contracts or other hassles. Email This email address is being protected from spambots. You need JavaScript enabled to view it. or call me 347-603-6442 for some ideas you will like.

Breaking: Goldman $5B Writeoff, Adtran Warning Point to Telco Capex Cut & Highly Distorted Earnings to Come
Goldman Sachs took a $5B writeoff this morning for the new tax law and Bloomberg expects $20B more from Citi. Verizon and AT&T have some of the same issues so may see similar. Do not believe U.S. earnings results in the next year or two because balance sheet adjustments may overwhelm underlying trends. Underneath it all, all these companies are extremely profitable and the tax changes will put them ahead. But the numbers the next 24 months could be highly misleading.     Verizon has $46B in "deferred long term liabilities" and $38B in poorly defined "other liabilities." Major adjustments - up or down - would surprise no one. Phil Cusick of Morgan did an analysis a while back from which I inferred Verizon would need writedowns. Since he hadn't said that explicitly, he got angry at me and cut me off his distro. (I'd like to be back. His analysis then and other times was on target.) I haven't trusted Verizon's financials for a decade because of unclear items like these liabilities and enormous goodwill. Former CFO Stephenson at AT&T has long been a wizard at delivering the numbers he wants to. 
      Suggestion: If you are putting money into this industry, closely follow analysis from John Hodulik and Craig Moffett. They are among the best at finding reality beneath the figures reported.
     Separately, Adtran pre-announced 4Q sales at $125M, $20M less than the last three quarters. Tom Stanton explained, "this quarter has been significantly impacted by a merger-related review and slowdown in the spending at a domestic Tier 1 customer. Our current understanding is that this review will be completed in 60 to 90 days, at which time capital plans will be finalized." Since Verizon doesn't have a large merger on the table, I thought that meant AT&T is slowing down spending. T has been telling the street the 4G build is mostly over and the 5G will be modest for at least another year or two, allowing a capex cut. I now hear it may be Century.
     Randall fired a few thousand people just before Christmas, part of the plan to cut 70,000 jobs that most people haven't noticed. AT&T decided a year ago or more there was no growth in telecom and they would have to cut back. One way to view the $85B Time Warner deal is that AT&T had no other good way to move that much money from the current business.
    Adtran is down 14% on the news and Calix 5% on fears they will also be affected. The details in the announcement are somewhat encouraging; Adtran expects to almost break even for the quarter despite a big sales disappointment. 
Both Adtran and Calix are doing well in key accounts, especially with their SDN software. I hope that continues; I'm in the process of selling them ads and I need the money.

*** Telefonica Deploys ASSIA’s GPON Expresse Solution
ClearView support for Wi-Fi and Expresse are now available for both DSL & GPON (ad) 

Misses (Hits next issue) 
"I make many mistakes," the Butler said.

  • Africa in 2017 to Have More Internet Users Than U.S. has people I made that prediction in 2014 working with Cisco data. Two very bad years mean Africa won't pass 320M until late 2018 or 2019.
  • Small cells, finally: U.S. carriers get going in 2016 Didn't happen. Telcos found CA and 4x4 MIMO gave them the needed capacity.
  • U.S. cable companies nearly always deliver close to promised speeds Time Warner committed major fraud including distorting the FCC/AskSam data I relied on.
  • Atif Mian of Princeton tweeted, Six Americans won the Nobel prize this year in various sciences. ALL of them immigrants. (Mian won in economics, I wrote,) (Actually, Mian hasn't won a Nobel.)
  • Gig LTE: Telstra 2016, ... "I believe that 'gig' LTE will deliver hundreds of megabits 90+% of the time" Updated estimate from Qualcomm is that 75-150 meg will be common, not too shabby. While Telstra announced they had Gig LTE in 2016, I discovered they actually didn't ship.

World First "1.2 Gig" Commercial Service At Ooredoo Qata
Limited customers, speed only occasionally reached. Qatar Airlines is the first to deploy Ooredoo's > 1 gigabit service, now being offered to selected customers in selected locations. C114, one of China's best tech pubs, is my source for the 1.2 gig speeds, confirmed by four other news publications. That the world leadership has left the U.S. and Europe is a major story, even if the details may be slight puffery.
   Over a year ago, I discovered that True in Thailand was one of the first in the world to deploy 4x4 MIMO across the country, matched in 2017 in many other networks. That's just starting to be matched in the U.S. and Europe. Five companies - Ericsson, Huawei, Nokia, Samsung, and ZTE - are the source of almost all the 4G & 5G networks today. They sell the same gear in Qatar, Myanmar, Berlin, & Chicago. Except for Verizon and 5G mmWave - coming in 2018 - the Western telcos are no longer ahead.
    The most advanced commercial wireless networks previously announced have been "up to 1 gigabit" using the Qualcomm 835 chip. Ooredoo is at 85% coverage in Myanmar, headed to 92%.

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." (ad)

Times of London: "YouTube a shop window for child abuse videos"
Google and Facebook continue to fail at blocking what society considers abusive. It may not be possible to do so without a truly massive team examining videos. Despite the German government threatening $millions/day in fines, government pressure everywhere, advertisers fleeing and thousands hired, the control systems are failing. 
     Alexi Mostrous, Head of Investigations at Murdoch's Times of London, just before yearend reports, "Child predators are using YouTube as a 'shop window' to showcase abused children before exchanging explicit footage and images with paedophiles around the world. One Brazilian paedophile posted a dozen videos of young girls to YouTube this month. Most were ten seconds long and showed the children standing silently, licking their lips or dancing. One showed a masked child aged about ten saying: "Hey guys I got new underwear."
     Each video was emblazoned with the paedophile's email address. When an undercover reporter made contact, the man boasted he had 315 gigabytes of material showing 'naked' children.
     Another alleged child abuser, calling himself Horny Pastor, was allowed to create a YouTube channel despite having a username that had been flagged to US and Canadian child-abuse authorities. 
    There's no easy answer to censoring clearly illegal content, like child pornography, without also censoring clearly legal speech, like the famous Vietnam photo of a napalmed young girl. Much more, including politician comment, a look at why it's almost impossible to do this right without an army of watchers, and why Google & Facebook are the most important censors in countries like mine

*** IoT Evolution Expo Orlando Jan 22-25
5 Collocated events, 80 Sessions, 250 Speakers, 2,000 Attendees & Unlimited opportunity (ad) Programmed by Carl Ford, one of the best

Vectoring Disaster: 26% Get Less Than 50 Megabits on a "100 Megabit" Network
50,000 refunds as complaints go up 160%. Much of the Australian network is second-rate. Belgacom got 80 megabits down to almost everyone. Most could get over 100 meg. Other telcos told me similar.
     When Malcolm Turnbull, now Prime Minister, asked me whether vectoring could deliver 100 megabits, I said yes. So I bear a small share of the responsibility for millions of Australians having a second-rate Internet for the next decade. (He heard the same from many others.)
     I (apparently) made an inaccurate assumption: that NBN would build the network to the well-known standards for vectoring. Vectoring generally delivers 100 megabits 500 meters and more than 50 megabits 800 meters. The figures below from the Australian Competition and Consumer Commission show ~50% are not getting the 100 megabits paid for and more than a quarter get less than half the speed. 3% do not even receive 25 megabits. There are rumors Deutsche Telekom is facing similar problems but no hard data. 
     Telstra is offering to compensate 42,000 customers for slow NBN speeds under pressure from the Australian Competition Commission. Optus is part-refunding almost 9,000 more. ACCC Chair Rod Sims is part of an international wave of regulators demanding telcos deliver what they advertise. OFCOM in England and BNetzA in Germany are cracking down. The British Advertising Standards Authority now requires companies to provide median speeds rather than the long denigrated "up to" speeds. 
     At Telstra, 26,497 (56%) of FTTN customers on the 100/40 Mbps plan could not receive 100/40 Mbps. Of those customers, 9,606 could not receive 50/20 Mbps. 6,352 (45%) of FTTN customers on a 50/20 Mbps plan could not receive 50/20 Mbps.
    At Optus, 5,430 (48%) of FTTN consumers on a 100/40 Mbps plan could not receive 100/40 Mbps, and 2,337 (21%) of those consumers could not receive 50/20 Mbps
    Switching from full fiber to vectored DSL has not solved the cost problem, unfortunately. The costs remain far too high. "We will continue to investigate others" writes Rod Sims, More data

*** The Broadband Forum thanks all the speakers and staff that made the BASE Summits a great event. Plan ahead for the BBF Quarterly Meeting in Athens, March 26-29 and the Summit alongside (psa) 

DNA Finland: Range of 3.5 GHz Massive MIMO = Range of 1.8 GHz Traditional
Adding more antennas makes adding cells unnecessary. I don't have cost details, but the innovative Finnish telco, DNA, suggests that using Massive MIMO in the 3.5 GHz spectrum could be very interesting. 
     Pal Zarandy writes, "DNA, one of the Finnish operators publicly shared their plans about using 3.5GHz with 5G massive MIMO, deployed on the existing 1800MHz macro site grid (i.e. no site densification needed, no Capex hikes), to push home broadband primarily in the outskirts of bigger cities, lower pop areas."
    This is consistent with what I heard from Softbank Japan, the first with a large deployment of Massive MIMO. Softbank told me they tuned for better performance at the cell edge rather than greater total speeds. Beamforming/Beamshaping gives the operator many ways to focus the signal as desired.
     Pal adds, "DNA has also a fixed-line business (broadband, cable TV), mainly in bigger cities. They have been pushing fully unlimited 4G MBB, now their mobile data volumes exceed their fixed-line volumes. In Europe the 3.5GHz spectrum auctions are coming up, operators are expected to acquire ~100MHz each (TDD)." 3.4-4.2 GHz will add 300 MHz or more in most countries, enough to double the spectrum available. More

Qatar Ooredoo Added to Massive MIMO Map, Dec 2017
Ooredoo is using Massive MIMO. Europe will fill in very rapidly, if only because DT is enthused and controls 7 Eastern European companies. I haven't heard much from Latin American leaders Telefonica and American Movil, but I expect that to change in 2018. Africa is the most intriguing possibility. Because they have almost no landlines South of the Sahara, African broadband will be overwhelmingly wireless. Capacity today is very limited, with caps generally 2 to 10 gigabytes. This is inadequate for the video Internet. 
    Massive MIMO is the natural solution, as Paulraj predicted in 2014. Softbank and others have demonstrated the technology works. Four executives at Huawei confirmed to me MM is ready for Africa. I infer from that the price makes sense and Huawei is ready to support the product there.


  • Harmonic, Ubiquiti, and Quantenna are good companies with excellent products. I strongly advise all but the best-financed professional investors to be cautious. The stock prices have been swinging wildly, far more than the 60 or so other companies whose news I track at Yahoo Finance. That usually means a few large investors are playing games with each other. The result is the stock moves aren't rational and there's no way to determine the right price. Harmonic makes some of the best encoders; Quantenna, probably the best Wi-Fi chips; Ubiquiti highly reliable inexpensive wireless gear.
  • Further warning: Blockchain probably won't live up to the hype but has real applications. Most Bitcoin buyers are suckers likely to get burnt. I had dinner with a fireman, a nice guy but highly unlikely to get rich on cryptocurrencies because he took a halfday course. If I understood more, I'd probably be even more skeptical.
  • Harold Ford, until recently a Comcast spokesman, left "Morgan Stanley after facing a human resources investigation into allegations of misconduct, a company spokeswoman confirmed. 'He has been terminated for conduct inconsistent with our values and in violation of our policies,' Verizon's HuffPo reports. Unlike Harvey Weinstein, there is no public evidence of rape or coercion. As the telcos and cablecos hide from reporters in D.C., we have to look for their proxies to find their positions. As far as I know, Mike Powell of the Cable Association, Jonathan Spalter of U.S. Telecom, & Meredith Attwell Baker of CTIA (wireless) never deviate from the positions of of their primary sponsors, AT&T, Comcast, & Verizon. Since Verizon and AT&T usually hide from reporters, I look to the mouthpieces for company positions. (I believe Ford, a liberal Democrat, took money indirectly from the cable guys.)
  • 4 on Huawei. COMMSDay: Huawei may yet have a chance at the Australian NBN after being blocked by the Aussie military. Malcolm Turnbull when he was Comms Minister told me he thought competition from Huawei would bring the cost far down.  ZDNet: "Huawei's $200 Honor 7X is the future of the smartphone industry;" Multiple news reports: AT&T will feature Huawei phones with $100M of advertising; Chinese press: Consumer chief Teng Hongfei has been detained with allegations of bribery. Every large company in the industry has faced corruption charges, including Huawei in Algeria in 2012. There is no evidence Huawei is in the same class as Siemens ($billions) or even the $hundreds of millions admitted by Alcatel, Ericsson, Telia and others.

Other recent stories:

Newsbreak: Kathy Brown's Leaving; Can the Internet Society Now Become Global, Independent, Democratic & Transparent?
Kathy Brown has announced her coming resignation as CEO of the Internet Society. It's her decision. She has strong support, continues to be one of the ablest policy advocates in D.C., and has the energy to work very hard. I've known her for many years to be one of the most decent executives in D.C. For the Internet Society, she brought the Internet Governance Forum back from a near-death moment to be a vibrant forum for discussing Internet issues. 
     Kudos are pouring in. Vint Cerf writes, "I would highlight Kathy's astute hiring of talent during her term as CEO and her expansion of the role of ISOC as a supplier of factual data and a defender of the Internet's openness." ... Internet pioneer Steve Crocker adds, "As past chairman of ICANN, I think Kathy's strong support of ICANN during our transition period should definitely be included.  This took significant energy and leadership and represented a high point for ISOC, for ICANN and for the entire Internet community." Nadira Alaraj emails, "Kathy's leadership is the approval of the Chapters Advisory Council." Angie Contreras applauds, "The creation of two very important groups: the Special Interest Group of Youngsters better known as the Youth Observatory and the Special Interest Group of Women." Renata Aquino Ribeiro remembers a personal touch. "In 2015, I went to the Public Forum in Marrakesh to discuss ICANN and gender balance, my 2nd ICANN meeting. Hours later I was outside the venue and she found me and complimented me for speaking out... Gratitude always."
     That said, my evaluation of ISOC under Brown was strongly negative. ISOC gets $30M/year in support through .org registrations. I believe it can and must be a vital voice for Internet users. That's why I care.

Warning: Boring policy discussion follows. 

ISOC under Brown has not lived up to claims to be Global, Independent, Multistakeholder/Democratic, Open, & Transparent. 
     ISOC leaders make eloquent speeches about how organizations like ITU should be "multistakeholder," open, and transparent. Our CEO has asserted ISOC is a "bottom-up multistakeholder" organization. Nearly all decisions are actually made top-down by strong executive leadership.
     I became disillusioned when ISOC supported U.S. policy in the great "North/South" split at the WCIT. The U.S. is refighting the cold war, not a productive policy.
     ISOC invariably chose to support the U.S. position over proposals from the emerging world with wide support. In particular, ISOC supported the U.S. on all important issues of access cost, such as reasonable royalties and multinationals paying taxes. The Africans told me the most important international factor in high costs is cartel-like pricing for backhaul and transit. ISOC opposed a proposal to do something about that. 
     The Internet was built and managed by the U.S. and other developed countries. That deserves enormous respect, especially to Vint Cerf who is actively disagreeing with what I'm writing. It's painful to dispute with Vint; developing TCP-IP was just his first major achievement. For the succeeding 40 years, he's traveled the world to get people connected. He's been personally generous. But the time has come to accept a changed situation and widen the decisionmaking.  
     The majority of Internet users are now in the rest of the world. The ROW, led by the BRICS, has been demanding a stronger role in key decisions. They believe, correctly, that many policies favored the countries in power.
     That the primary dispute was over power and influence, not human rights, was made clear to me by U.S. Delegation lead Larry Strickling. I asked why the U.S. was fighting so hard on trivial issues. Larry looked at me and asked, "Dave, do you want Russia and China to rule the Internet?" 
     While I believe Cerf and Brown were motivated by a belief in human rights, the 14 members of the U.S. delegation from three-letter agencies (NSA, CIA, HSA, DOD) were not there to protect freedom of speech.
      My answer, I realized later, should have been, "Of course they should have a role. The system is unsustainable if half the Internet users know they are excluded."
      History so far is proving me right. The BRICS, blocked at the ITU, are setting up a strong, alternative system, including making 40% of the contributions to 3GPP, the wireless group that is the key standards body. Nothing can be approved there without Chinese concurrence. Tim Cook of Apple kowtowed to China at the Wuzhan World Internet Conference, joined by Sundar Pichai of Google. I have more evidence from the BRICS meeting and the Belt and Road. Putin has started the process for a complete alternate root.
      I find the result of refusing to compromise is that the U.S. has much less say in the major decisions about the Internet, Internet commerce, and Internet taxes. The Chinese and Russians at WCIT were mostly concerned with the Internet in their own countries and I believe were happy to compromise on everything else. Blocked at ITU and ICANN, they feel they must build their own system. 
     ISOC refuses to apply our own principles. ISOC is not global: Easily half the Internet is essentially not represented within ISOC, despite our claims." There are no chapters - or board representation - from China, 1/3rd of the Internet. I called for a "Nixon goes to China" gesture by Brown. After years of discussion, the situation has not improved.
     Perhaps inadvertently, we actively discourage membership by those who disagree with our limited point of view. On our home page, we proclaim "Internet Society chapters are communities of like-minded people." I certainly am not "like-minded" with the leadership of the Internet Society; I often, but not always, agree with the Global South on issues opposed by ISOC leadership.
     We refuse to reveal our donors, acting like a DC Astroturf shop. I raise the issue because I have several examples of ISOC policy being influenced by donors. 
     We have a Chapters Committee but its most important recommendations are not supported. An example was a proposal for the chapters to allocate the huge sum of 3% of the budget (with fiscal controls) and make some local decisions, such as which policy events to attend. It received unanimous approval from the chapters committee. Initially, there was very strong board support. 
     The very modest Chapters proposal was completely rejected in a closed meeting led by a senior staffer. 
     I was severely flamed for these comments, an effective demonstration of how dissent is discouraged in ISOC. No one has questioned any of my facts. Brown's complete statement and far too much more, 

*Reply "subscribe" to be added, "un" to be dropped
Volume 18, #10  Dec 28, 2017

Dec 17

Net Policy News

Reply "subscribe" to be added, "un" to be dropped

Tim Cook of Apple kowtows in Wuzhan, China, joined by Sundar Pichai of Google. China has nearly three times as many Internet users as the U.S. No U.S. executive except Sergei Brin is willing to give up the Chinese market. 4 of the 10 largest Internet companies are Chinese. Nothing gets approved in wireless without Chinese approval. 
     This is the world today, like it or not.

Vladimir Putin’s Russia and the BRICS are building an alternative Internet root after years of talk. At the end of this issue is the first in depth look in English. 
     The Russians fear that the U.S. and allies will extend the sanctions and have ICANN cut off Russia. I think that unlikely but not impossible. 14 of 18 ICANN board members are from the U.S. and allies. None come from over half the Internet, including China & Russia, despite the ICANN CEO promising the Chinese “a seat at the table.” 
     This actually could be fine. Columbia Prof Eli Noam reminds us that the Internet is a “network of networks.” It’s not a monolith that can be split and destroyed. Noam and ICANN CEO Chehade agree that with robust connection, this could work well. 
     The smart policy for the U.S. and ICANN is to cooperate to make things work rather than re-fighting the cold war. It far more likely the U.S. will continue an unproductive fight over ICANN, which doesn’t really do very much.

30 second editorial: $20B Bharti just bought a Rwanda network from $6B Millicom, subject to regulatory approval. There’s enough money in the deal they would build crucial infrastructure if if the regulator is strong. #Rightmove 

"A Brain Drain & An Empty Shell," If Broadcom Buys Qualcomm
Will they ever again produce a chip like the new Snapdragon 845? The 835 is the best mobile chip in the world, dominating the market in top end smartphones. The newly announced 845, due in June, is reported to be 25%-30% faster. It will be produced on the same 10 nm process at TSMC. The 845 is "A significant upgrade ­ to get 25-30 percent performance in the same general process node is huge," analyst Jim McGregor tells Rick Merritt in EE Times. It's designed for speeds into the gigabits and ultra HD. 
     Jon Peddie's comment in the headline probably goes too far, but the cutbacks at Broadcom foretell problems ahead. Rick is one of the two or three best tech reporters working today. He wouldn't have run the quote if he hadn't heard from friends who worked at Broadcom. 
     Hock Tan has convinced some very sharp people on Wall Street that he is an exceptional manager who can fire a slew of engineers (~30%) and continue to produce world-class products. We won't know for sure whether Tan can produce near-miracles for several years, but a least one significant product is delayed. Broadcom's 35b DSL chip was promised for 2016 and then early 2018; Deutsche Telecom recently said it won't be ready for deployment until 2019.

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." (ad) 

4G Speeds: Reliance Jio India 22 Megabits; T-Mobile U.S. 18; Verizon U.S. 15
Slowly, the U.S. has dropped to #57 in the world. Reliance Jio has a world-class network built in the last three years. Verizon had a world-leading network in 2009-2012, the first large LTE network in the world. 
     Singapore, South Korea, and Norway are the fastest. Eastern Europe does surprisingly well, presumably because their networks are more recent. U.S. 4G speeds are half those in Canada and Spain. We are behind Russia, Brazil, Tunisia and dozens of others. Map and lots of data

*** Telefonica Deploys ASSIA’s GPON Expresse Solution
ClearView support for Wi-Fi and Expresse are now available for both DSL & GPON (ad) 

Gig Of Data Too Effing Expensive For Billions
From Nicaragua to Zambia, a basic 1 gig/month connection is too expensive for half the population. In Vietnam and Pakistan, 1 gig costs less than 2% of the average monthly income, the A4AI goal for affordability. In Nicaragua, a comparable figure is 8% ; in Zambia, 12%. 
   The key question is how to change that. One gig barely allows a couple of hours of video so a real minimum should be higher. The technology is making lower prices/higher caps possible. Verizon estimates the cost of delivering wireless data is going down 40%. So it's not a surprise many of these prices have come down. 
    As Reliance Jio is proving in India, new technology is ready for emerging nations as well.  Take a look at (not . com) for some ideas on the right technology
    The most important international policy problem is high backhaul/transit costs, which are often cartel-like. They often are 10-40X higher than in the developed world and add $5-$10 to a robust Internet package. Americans should be embarrassed we are blocking solutions.
   Most of the other policy choices are best made domestically. Almost always the people on the ground know more than "policy" people in DC or Geneva. Nice map and more

Four Antennas Good, Two Antennas Bad. Eight Antennas  50% Better?
T-Mobile Netherlands and True Thailand are using 8 antennas for FDD spectrum, joining Kuwait in choosing Huawei gear. 8 antennas will rarely deliver twice the throughput of four antennas but will often deliver a 50% increase in capacity. 32/64 antenna Massive MIMO has been chosen by BT, DT, & FT for higher frequencies running TDD.
    32 antennas for 700 MHz or 1800 MHz spectrum would be too large to be practical most places. In those frequencies, a very senior engineer recommends "Don't go beyond 8 antennas for FDD for now." 
    In addition, most frequencies below 2000 MHz are dedicated to FDD transmission, with upstream and downstream in separate bands. This produces problems with upstream signaling, often a major issue for lower frequencies.

*** The Broadband Forum thanks all the speakers and staff that made the BASE Summits a great event. Plan ahead for the BBF Quarterly Meeting in Athens, March 26-29 and the Summit alongside (psa) 

Björnson Publishes Massive MIMO Networks: Spectral, Energy, and Hardware Efficiency
Professor Björnson's book comes at an auspicious moment. British Telecom, Deutsche Telekom, & France Telecom/Orange recently committed to large deployments of Massive the recent Huawei Mobile Broadband Forum. The book contends, "A lot of classic 'wisdom' about Massive MIMO is questionable because it is based on too simplistic system models."
   The book presents alternatives over 516 pages, rigorous, loaded with lemmas, corollaries, and math. For a professional monograph, the writing style is clear.
    Stanford Professor Goldsmith has told me a deep understanding of capabilities of systems with multiple receivers and transmitters is one of the most important problems in communications. Shannon's Law is well understood for a single transmitter and receiver, but many systems now are more complicated.

*** IoT Evolution Expo Orlando Jan 22-25
5 Collocated events, 80 Sessions, 250 Speakers, 2,000 Attendees & Unlimited opportunity (ad) Programmed by Carl Ford, one of the best

Net Policy News (Gfast will be back next issue.)
The world needs a good news source on Internet and telecom policy. I hope to create one.

Professor Noam's "Many Internets"
Until about 2010, everyone agreed the Net was a "network of networks," not a monolithic entity. There was a central authority, ICANN, keeping track of domain names, but that was a minor administrative function. 
    Columbia Professor Noam suggests we might be better off accepting that some nations or groups might want to organize their networks differently. It's easy to see demand for an Internet with much more effective filters against material some think harmful to children. (Any 10 year old can easily find porn today. Many do.)
     Internet translation is getting better very quickly. You might want an "Internet" that translates everything into your language. Google Chrome translation isn't perfect but I was able to research most of this story on Russian language sites. With a few more years progress, I might welcome an alternate that brings me everything in English, including caching for better performance.
     De facto, Internet news is already split, as hundreds of millions only get their news from Facebook. Google AMP pages, including for news, also favor selected parts of the net
     Centralizing the DNS doesn't prevent censorship, as the Chinese have demonstrated. There are many Jewish and Muslim fundamentalists who want to block what they consider blasphemy and limit free speech. See . More from Noam

Russia Orders Alternate Root Internet System
It's actually practical and not necessarily a problem.The Security Council of the Russian Federation, headed by Vladimir Putin, has ordered the "government to develop an independent internet infrastructure for BRICS nations, which would continue to work in the event of global internet malfunctions ... This system would be used by countries of the BRICS bloc – Brazil, Russia, India, China and South Africa." RT
     Columbia University Professor Eli Noam and then ICANN CEO Fadi Chehadé have both said such a system is perfectly practical as long as there is robust interconnection.
     Actually, the battle over ICANN and domain names is essentially symbolic. Managing the DNS is a relatively insignificant task, more clerical than governing. ICANN Chair Steve Crocker pointed out they had very little to do with policy. 
     Some will claim this is about blocking free speech but that's rhetoric. Russia doesn't need to fiddle with the DNS for censorship, as the Chinese have demonstrated. The wonders of the Internet will continue so long as the resulting nets" are robustly connected. The ICANN and U.S. policy goal should be to help create that system for interconnection. 
    I expect contentions that “The Russians are taking over our Internet” and “They are splitting the Internet.” The Internet is a “Network of Networks.” It is not a monolith so what would “splitting” it mean or do? 
    After the WCIT, China realized that ICANN and the DNS are side issues not worth bothering about. They have been building alternate institutions including the World Internet Summit in Wuzhan and the BRICs conferences.  The Chinese have put their main work where decisions that matter are made. Wireless standards are set by 3GPP, where nothing can be approved without China's consent.
     The American battle at ITU is proving to be a historic mistake.
Why does Russia want an independent Internet?
    They fear that Western sanctions on Russia could cripple the Russian Net. Communications minister, Nikolay Nikiforov, worries about, "a scenario where our esteemed partners would suddenly decide to disconnect us from the internet." I think that's highly unlikely but Nikiforov points out, “Recently, Russia is being addressed in a language of unilateral sanctions: first, our credit cards are being cut off; then the European Parliament says that they’ll disconnect us from SWIFT."
     It makes sense for the Russians to be prepared for such a contingency as the Cold War has been warming up on both sides. "Britain's top military chief Air Chief Marshal Sir Stuart Peach just made headlines warning Russian subs "could CRIPPLE Britain by cutting undefended undersea internet cables." Much more

ICANN Continues Excluding Russia & China From the Board
No wonder Russia wants an alternate root. Three years ago, ICANN CEO Fadi Chehadé promised "a seat at the table" to Chinese Premier Li. ICANN welched and this year added two more Americans. 
    Almost all the ICANN board is from the U.S. and close allies; only about 4 of the 18 board members are from countries on the other side of the North/South divide in Internet policy.  Claiming ICANN represents the Global Internet is inappropriate. China is 1/3rd of the Internet but has no representation on the board. 
     I know many of the board members. They are all basically honorable but generally share a strong opinion on North-South issues. 
     Larry Strickling of the U.S. government knew just what he was doing with the IANA transition. He handed over to a board with similar positions as the U.S. government.  
     "The system is unsustainable while it excludes half the world," I have been saying since 2012. More, including the transcript of Fadi's statements,

Sorry, Ajit Pai: Smaller Telcos Did Not Reduce Investment After NN Ruling
Pai justifies his NN choice with the claim, "The impact has been particularly serious for smaller Internet service providers." #wrong (Actually, NN has minimal effects on investment, up or down, I’m convinced. Competition, new technology, customer demand and similar are far more important.)
     The two largest suppliers to “smaller ISPs” saw sales go up. Adtran's sales the most recent nine months were $540M, up from $473M the year before. 2016 was $636M, 2015 $600M. Calix the last nine months sold $372M, up from $327M. The full year 2016 was $459M, up from $407M in 2015. Clearfield, a supplier of fiber optic gear, was up 8% in sales in the smaller ISPs.
      There is nothing in the data from others that suggests an alternate trend. Anyone could have found this data in a few minutes from the company quarterly reports.
     The results in larger companies are ambiguous. I can "prove" capex went up or went down by selecting the right data. The four largest companies' capex - two/thirds of the total - went up from $52.7B in 2015 to $55.7B in 2016. The result remains positive after making sensible adjustments for mergers and acquisitions. That's as close to "proving" that NN led to increased spending as the facts chosen to prove the opposite.
      Actually, whether capex went up or down in 2016 tells us almost nothing about the choice on neutrality. Everyone knows a single datapoint could be random or due to other causes. Much more, including the source of the errors

Elders Bearing Witness: Vint, Timbl, & Many More
Vint Cerf, Tim Berners-Lee, Steve Wozniak and more than a dozen true Internet pioneers wrote Congress to protect Neutrality. The best Congress money can buy didn't listen but I wanted to reproduce their letter. 
    I hope they are wrong believing "is an imminent threat to the Internet we worked so hard to create." My take is the impact will be moderate in the short run. 
From the letter: 
We are the pioneers and technologists who created and now operate the Internet, and some of the innovators and business people who, like many others, depend on it for our livelihood. ... The FCC’s proposed Order is based on a flawed and factually inaccurate understanding of Internet technology. These flaws and inaccuracies were documented in detail in a 43-page-long joint comment signed by over 200 of the most prominent Internet pioneers and engineers and submitted to the FCC on July 17, 2017.
     Despite this comment, the FCC did not correct its misunderstandings, but instead premised the proposed Order on the very technical flaws the comment explained. The technically-incorrect proposed Order ... More, including the full list,

*Reply "subscribe" to be added, "un" to be dropped
Volume 18, #9  Dec 16, 2017

Nov 30

Reply "subscribe" to be added, "un" to be dropped

Breaking: Verizon announced 5G to 30M homes. Reporters from the Wall Street Journal to the Albanian Times treated this as big news. I had the story May 2, if I may brag a little. “The decision had been made to go to a one-third to one-half the United States.” Key takeaways: "Every variable tested during the trials came out better than assumed," including 600 meter median reach. 5G mmWave is so cheap capex won't go up.

On the contrary, BT, FT, & DT recently said they would go slow on 5G mmWave (2022-2024 for most.) They already cover nearly their entire country with broadband; Verizon only about one quarter of the U.S. VZ has incentive to build out of market; the Europeans don’t. More below.

London: Princess Leia was here at the Huawei MBBF. Huawei hosted a 5G demo of a woman suspended in space. It wasn’t actually Leia, but you almost expected the highly realistic 3D rendered figure to start saying “Help me Obi-Wan Kenobi. You’re our only hope.” (It also was on a special 2D screen rather than a hologram floating in mid-air.)

A few feet away, attendees were invited to play goalie is a small soccer cage. Your eyes would be covered by a VR rig, through which you’d see the actual soccer ball be kicked at you. Most of the time, the VR-enabled player blocked the shot. 

BT CEO Gavin Patterson, DT CTO Bruno Jacobfeuerborn, and FT/Orange SVP Arnaud Vamparys all expressed hesitation about large 5G builds. “There’s no business case,” all three declared, although they hope that will change. In 2019-2020, they will promote their 5G work, but intend to go slow spending until 2022-2024.

Patterson, Jacobfeuerborn, and Vamparys all endorsed Massive MIMO and said it would be used on their networks. TDD Massive MIMO is ready to triple LTE performance at a cost so low Huawei expects to sell it in sub Saharan Africa. That will add England, France, and Germany to the Massive MIMO map. has more reporting on Massive than any other site on the web. 

*** Telefonica Deploys ASSIA’s GPON Expresse Solution
ClearView support for Wi-Fi and Expresse are now available for both DSL & GPON (ad) 

Verizon 5G: "Every variable tested during the trials came out better than assumed."
A gigabit median at 600 meters. A year ago, most experts expected mmWave to be limited to 150-250 meters. Earlier this year, Verizon mentioned good results at 400 meters. Now, Verizon is often seeing excellent results to 600 meters and further. A gigabit is common. 

The result is that far fewer new cells are required.  The cost is coming in so low Verizon does not expect to raise their capital spending. "Management stated that overall capital expenditures would not change substantially during the 5G rollout as internal resources are shifted from the 4G to the 5G platform."

It will be fixed only in 2018, with Verizon promising to support mobile as soon as the technology is available. Intel and Qualcomm expect 5G phones in production in 2019.

Verizon will do five cities late in 2018, starting with Sacramento, an AT&T town. They promise competitive prices, which will be required to reach their goal of 30% market share. Both Verizon and AT&T expect whoever covers a city first will win 20% to 40% away from cable or DSL. It's a land grab situation, which I expect will inspire AT&T to move fast as well. 

Verizon will probably choose two vendors among Ericsson, Nokia, and Samsung. Samsung had few 4G builds outside of Asia, but has been investing heavily for half a decade.

"The Business Case Isn't There." Why BT, China Telecom, DT, FT, Telefonica Slowing on mmWave
At Huawei's London MBBF, BT CEO Gavin Patterson went first, saying, "The business case isn't there." DT CTO Bruno Jacobfeuerborn, adding "For 5G, Deutsche Telekom still has no business models."  FT/Orange SVP Arnaud Vamparys followed with a second agreement.

3,000 miles away, Verizon was finalizing the announcement that they would rush to deploy ~30M homes of 5G mmWave. 

Verizon landlines cover only about 1/4th of the U.S. They need an offering to compete with cable and DSL. They believe the new one gig wireless service will win 20-40% of the landline customers. AT&T is thinking the same way and will probably invest to keep up with Verizon.  

The Europeans have a landline offering throughout their territory.  All of these companies will begin deploying in 2019 & 2020 because it's a political necessity. Expect plenty of hype and pr. They also need to train their people and develop systems for when they do build. 

But the current plan of each company is to go slow until 2022-2024.

England, France, and Germany Promise Massive MIMO
BT CEO Gavin Patterson, DT CTO Bruno Jacobfeuerborn, and FT/Orange SVP Arnaud Vamparys all endorsed Massive MIMO at Huawei's London MBBF.  They spoke one after the other at Huawei's Global Mobile Broadband Forum in London on November 14. 

BT has a Massive MIMO rig in active testing.They only have 15 MHz of TDD 2.6 spectrum, although they have 50 MHz of FDD spectrum. They are actively watching the development of FDD MM, especially as the 3.5 auction continues to be delayed. They could add roughly 70% to network capacity if OFCOM allowed them to switch the FDD spectrum to TDD. They support MM research at Bristol University.

Deutsche Telekom is trialing 3.5GHz Massive MIMO. DT has so much LTE capacity they are promoting hybrid boxes for DSL and LTE. They have carefully not promised any upgrades to the landline network to 20% of the population. That suggests they intend to do what AT&T & Verizon are planning: turn off the wired network in a large share of the country.

Orange/FT MM could be particularly important in Orange's dozen African networks. Because Africa has few wires, seriously expanded wireless capacity is the only way for a decent video-grade Internet.

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." link (ad) 

Iliad Free Already Bringing Italian Down Prices
3 to 4 good. 4 to 3 bad. Xavier Niel's Free Mobile France offers a 100-gigabyte cap on their 20 euro service, which Jennie bought in Paris. For 20 euros, Telecom Italia offers 5-10  gigabytes. The anecdotal evidence is strong that prices move 10-20% when a market goes from 4-3 or 3-4 primary carriers.

Like Ambani's Reliance Jio in India, Xavi is building a state-of-the-art network with costs much lower than the existing players. Vittorio Colao of Vodafone knows he'll have to respond. "When you have a big warship and the pirates are approaching, it makes sense to send out the commandos in the speed boats.” 

With government-backed Open Fiber also building, Italy will eventually catch up to Europe.

*** The Broadband Forum thanks all the speakers and staff that made the BASE Summits a great event. Plan ahead for the BBF Quarterly Meeting in Athens, March 26-29 and the Summit alongside (psa) 

10,000 MM Cell Sites Shipped - All TDD, Not FDD
Zhu Huimin indicates Huawei is moving to high volume production. I once suggested, "Don't believe any equipment claims until 10,000 are deployed." Most problems, including those from heavy use, will appear by then. By that test, Massive MIMO - at least the TDD variety - is now solid. The first announcements of Massive MIMO commercial use were just after a year ago. Softbank Japan was the pioneer; I broke the news that China Telecom was also buying thousands. Although Huawei does have a 32x32 antenna FDD Massive MIMO, her recommendation is to stick with 8x8 in FDD spectrum.

There is no doubt that TDD MM works and will rapidly deploy around the world. Two Huawei executives confirmed to me Massive MIMO is ready for Africa. The need is great; there are very few landlines so almost all the Internet connections are mobile.

All are TDD; FDD is only in trials, with major problems still to solve. Massive MIMO is ready for hockey stick growth.

*** IoT Evolution Expo Orlando Jan 22-25
5 Collocated events, 80 Sessions, 250 Speakers, 2,000 Attendees & Unlimited opportunity (ad) Programmed by Carl Ford, one of the best

U,S. Q3: Cable +540K, Telcos -155K. Telcos clobbered across > 1/2 the countr
"In the first three quarters of 2017, cable companies added about 2,000,000 broadband subscribers, while Telcos lost about 430,000 subscribers." Bruce Leichtman. Ten years of telco dis-investment across half the United States is allowing cablecos to rise from 55% of the market to 64%. They are pulling further away every quarter. 

AT&T's stock went down $13B on they day they warned about disappointing numbers. If they had spent even a fraction of $13B on investment, their numbers would have been fine. Verizon's investment % has gone down for several years; this quarter, they lost 10,000  broadband customers.

Frontier lost 63,000 subscribers, The stock has gone down 85% this year to a level that implies a serious bankruptcy risk. That is unlikely before 2020 because they don't have large debts coming due. They have been dis-investing for years, with capex less than depreciation. Centurylink lost 101,000 for the same reason. 

I only have one choice for a decent Internet connection, Time Warner cable. Verizon can only supply 3 megabits to my building in Manhattan. About half of the U.S. has only one source, the cableco. Cable prices have gone up 30% during a period the actual costs of broadband have been flat to down. That's a market failure.

Gfast News
I’m still working through remarkable presentations from the Broadband Forum events. Michael Weissman, Bernd Hesse and team did a remarkable job choosing the speakers.

Deutsche Telecom: 35b Supervectoring Delayed to 2019
Broadcom is now over 3 years late. DT briefed German reporters after their financial call and revealed 35b was now delayed until 2019. 35b should deliver 200+ meg downloads 500-600 meters, a crucial tool for DT, which is losing share to cable. Cable now covers about 70% of Germany and is expanding. DT now only offers 50-100 megabit DSL while cable is often 400 megabits, going to a gigabit. 

The problem is software; the hardware is shipping and supposedly will work. DT says 35b is not ready to turn on. Broadcom in 2015 said 35b was in "production" in the press release below. Alcatel in early 2016 said to expect complete systems very soon. "35g is very similar to 17a so there should be little delay."

Broadcom's problems are leading major telcos and vendors to have a plan B, using Sckipio DT itself is planning extensive deployments in 2019, mostly in apartment buildings.

Gigabit 100 Meters - Unless the Wires are Lousy
Speeds are fine, "Unless there's a line problem." I've been reporting for three years that ~10% of lines have problems. In the chart by Rami Verbin of Sckipio, he finds goes ~130 meters on good lines. Poor lines have about half the reach. 

His chart roughly matches the reports from Swisscom, Belgacom, and England for both & vectored DSL. The 10% with problems can cause the majority of the line-related complaints to support. The angry customers drive up cost.

Rami's solution to reach the gigabit is bonding, supported on the Sckipio chips. Verbin made some additional points:

  • 4 gigabits is possible by bonding two decent 2 gigabit lines.
  • Even in a service from remote cabinets, ~25% are close enough to get a full gigabit."
  • cDTA and iDTA are practical ways to deliver much higher upstream by switching some bandwidth from downstream to upstream only when needed.
  • 35B will probably be similar but Deutsche Telecom doesn't expect to deploy until 2019.
*** The Broadband Forum thanks all the speakers and staff that made the BASE Summits a great event. Plan ahead for the BBF Quarterly Meeting in Athens, March 26-29 and the Summit alongside (psa) 
AT&T Wants Coax 2-5 Gigabit Very Soon.
AT&T faces intense competition from cable, talking about 10 gigabits in both directions. (Cable will only be 1 gig down, ~100 meg up, until ~2021.) AT&T wants something to brag about as well.

AT&T gained millions of lines of coax as part of the DirecTV deal. Alcatel and Huawei are leading the development of G.mgfast. That uses 424 MHz, full duplex, to achieve ~2.5 gigabits in both directions. The reach on telco twisted pair is only about 30 meters. On coax, those speeds can probably extend far enough to service most apartment buildings. Using 848 MHz, speeds can reach 5 gigabits. The ITU standards group has been aiming for 2019-2020 for G.mgfast, too slow for AT&T's marketers. David Titus wants a high-speed standard for coax "early in 2018." He believes that is "doable."

PLDT Puts Philippines on the Map - Slowly
"The telecommunications duopoly is about to end,” President Duterte promises. expecting the competition to improve services and lower prices. He invited the Chinese to be the third player in the Philippines in a meeting with Chinese Premier Li Keqiang.

Incumbent PLDT is responding by connecting 100 buildings with and 35b Super-vectoring and some other places. PLDT is buying the gear from Huawei. Foreign-owned PLDT has a near monopoly on the Philippine Internet (85%) and minimal regulation. Their capex has been going down despite a dismal Philippine Internet. Filipinos have the slowest Internet in Asia, as you can see in the chart below from Akamai.

China Telecom is expanding internationally, including putting $6B on the table for bankrupt Oi Brazil. China Telecom has fiber to "essentially all of Shanghai and many other cities." Working with the Russian Railways, they have fiber across the "Belt and Road" and on through Europe. China Mobile has major investments in Thailand and Pakistan. Without a doubt, either could build throughout the Philippines in a very short time. Think Gig LTE to 85-95% of the country in 18 months, with Massive MIMO in the densest areas.

Ending Net Neutrality: Copps is Wrong "It Will End the Internet As We Know It." Pai is Wrong It Will Drive Up Investment

We've all heard too much about NN, which I've been reporting for 20 years. I support it because I don't want Randall Stephenson of AT&T deciding what I should watch on TV. The long-run effect is negative. 

The claims from some people who agree with me are ridiculous.  "According to former FCC commissioner Michael Copps, ending net neutrality will end the Internet as we know it." Michael knows I respect him, but this is ridiculous. Equally unlikely in Pai's belief this will significantly raise investment. His key evidence is that AT&T cut investment, which he attributes to Wheeler’s NN. That would be true iff AT&T has a time machine; well before anyone expected Wheeler to be strong, they had told Wall Street the cuts were coming. ("We pretty much finished the LTE build so we will reduce spending.")

The only major change likely in the next few years is some video will go up in price. The actual fights are about the cost of connecting for high volumes of data, mostly video. No one is going to redirect you and me from the N.Y. Times to Breitbart News. 

Pai could discover his error by listening to AT&T CFO John Stephens next Tuesday at UBS. CEO Randall claimed he would raise his $22B capex if NN died. 10% of that would be about $2B. He also said he would invest another $billion if he got the tax cut. 

If they were going to live up to this, Stephens would almost be required by SEC rules to tell Wall Street capex was headed up to ~$25B. He won't, but Pai is so set in his beliefs he won't be able to believe that.


Nov 6 News: Deutsche Wants a Gigabit, Finally Realizes 50 Meg Isn't Enough,1.6 Gig in Sckipio-Calix Test, Reverse Power 4 Port DSLAM for Australia, Australia Makes it Official: to Million Plus, 2 Bonded 212 Lines = 3 Gigabits

*Reply "subscribe" to be added, "un" to be dropped

Angelina Jolie, Ashley Judd and dozens of others have exposed Harvey Weinstein. Every industry, including ours, has monsters like that, rarely stopped.
    I choose instead to highlight some women who are driving us forward. Nicola Palmer and Sanyogita Sangupta lead Verizon’s 5G effort, the most successful in the world. They will spend $billions and have already put hundreds of cells in the field  
     Verizon jumped ahead when almost all others held back. Charla Rath persuaded the FCC to offer gigahertz of mmWave spectrum three years before the official WRC plan. They supported the NYU work to make a more accurate channel model.
     With allies, VZ created their own 5G standard in 2015 because they didn’t want to wait for 3GPP.  The system commercial in 2018 and will do mobile in 2019 if the handsets are ready.  
     “The race is on” between Verizon and AT&T to be first in each market.10%-20% of people almost everywhere hate their broadband provider and likely to switch to mmWave. The U.S. build is going fast because the two giants expect to find a large market replacing landlines outside of their existing region.
     If they are first in a land grab.

The Shadow CTO 
Our new site for decision support. The Wireless Troika - More antennas, More spectrum, More cells - is a good way to look at your choices. (Today, 4x4 MIMO antennas and CA to use more spectrum you own. Massive MIMO is working beyond that.) There's a Reality Check on (deploying at dozens, current speeds ~750 meg per AT&T, second gen chips in 2018 easily a gigabit.)
I thought it might be interesting to consider running a network.

Why AT&T - Time Warner is on Hold
AT&T won't make the concessions the Justice Department staff are demanding. One report is they want a spin-off of HBO; another says it would be CNN. On BellSouth and DirecTV, AT&T bamboozled DC and made almost no real concessions. With Randall Stephenson being one of the very few to publicly support Donald Trump, they expected no problems from the government. The spinoffs mentioned do not make a big difference in the deal. AT&T just doesn't want to give anything. 
Two notes: That AT&T-TW doesn't come under antitrust because it is a vertical merger is bs on the law and the facts. The movie studios were forced to divest the movie theaters in a textbook case, a similar "creation-distribution" situation. AT&T is prominently saying "We are an entertainment company," after the DIrecTV purchase and their video distribution. They are even producing new TV series. So there is a direct conflict.
An insider thinks Trump will insist that AT&T muzzle his critics at CNN. Courage, Randall.

Why Sprint & T-Mobile Didn't Merge
The common story is that testosterone-fueled Masayoshi Son wants power so much he killed the $60B Sprint-T-Mobile deal. A deeper look suggests that DT wasn't willing to pay the price Masa-san wanted to give up control. DT low-balled the offer, thinking Sprint had no choice but to give in. (Craig Moffett's analysis allowed an inference that without the deal Sprint might be insolvent soon.) Masa-san is a brilliant trickster who may yet sell to T-Mobile. The reduced competition would divert perhaps $20B/year from consumers to telcos.

Wireless Network Design: The Troika
The three basic tools for capacity more antennas, more spectrum and more cells. 
     Adding antennas is usually the cheapest choice in 2018. Nearly every telco is going to four (4x4 MIMO.) Adding spectrum is very attractive when you have some unused but expensive if you have to buy it. Adding cells was the approach before LTE-A in 2014 permitted aggregation. Since then, tower building has almost stopped in the developed world. Small cells in targeted areas are coming into fashion, but their numbers will be modest until mmWave.
     Any upgrade should include 256 QAM for more bits per Hz. More

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." (ad) 

5G Millimeter Wave Reality Check
Verizon & Korea are building. NTT DOCOMO & China Mobile are close. Many others are mostly holding back, doing little until 2021-2023.
     2019 is likely for mobile phones, Qualcomm promises chips for retail phones, as TSMC ramps 7 nm process to produce the chips. I expect the early phones will be power hungry and battery short. 
     Latency will be 5-10 milliseconds to the cloud controller intelligence. 1-2 milliseconds will be measured from the tower to the phone, but that's meaningless until and unless intelligence moves to the edge. That's likely very rare for a decade because there's no apparent market for 1 ms. 
     Most "use cases" are bogus according to experts in connected cars, virtual reality, and telemedicine. Gerhard Fettweis's Tactile Internet is inspiring but years or decades away most places. More

*** Telefonica Deploys ASSIA’s GPON Expresse Solution
ClearView support for Wi-Fi and Expresse are now available for both DSL & GPON (ad) Reality Check
It works and is deployed in dozens of places, including British Telecom and AT&T.
     T reports typical speeds of 750 megabits, up and down combined. The highest speeds go less than 150 meters, which is fine for fiber to the basement. In 2018, 212 MHz chips will double that speed. Plans include 10M homes passed by BT and a million at nbn Australia. Much more, including details of vendors and likely 2018 changes

*** Columbia CITI Nov 17 State of Telecom
Eli Noam and Bob Atkinson have the very best in world telecom to the university for this annual event. It was unmissable for anyone who cared about policy. This year, it's going online, (psa)

Qualcomm claims $trillions of 5G benefits from applications that don't need 5G
Steve Mollenkopf's creative engineers produce many of the world's best chips. It's time for him to stop promoting a bogus claim "5G Expected to Create 22 Million Jobs, Produce Up to $12.3 Trillion of Goods and Services by 2035." They offer no evidence or logic that leads to that conclusion. They hired IHS Markit to do a "study" and Berkeley Professor David Teece to sign off on it. 
    The substance of the study is a laundry list of future services, from drones to telemedicine to IoT to VR to Smart Grid. (They provide no source for their estimates of the values of each, but added together they do come to enormous figures.)
     The problem: Nearly none of this requires 5G and the report gives no reason to believe their value-add should be attributed to 5G. Nearly all of them work just fine on LTE or Wi-Fi and do not require 5G 

Verizon's 5G Gear From Texas! Startup
Farooq Khan says Phazr, his small company near Dallas, can compete with the giants for the millions of 5G small cells soon to deploy. Verizon is listening and is testing the Phazr equipment. Their target was to be ready by the end of 2017.
     Phazr's base station is about two feet high and a foot wide. That's large enough for 384 antennas per sector, three sectors per cell. That's massive, massive MIMO. It's made possible by the very small size of mmWave antennas.
     Phazr's system uses mmWave for the download but ordinary spectrum below 6 GHz for the upload. This has the advantage of requiring much less power in the mobile phone, producing less heat.

Next Gen Wireless Nov 29. Worth the trip.
Mike Dano and Fierce have put together a one-day event with some of the most interesting people in U.S. wireless. These are the people building the networks, not the salesmen and political silver tongues who plague most events. If I have the travel budget, I'll go myself. The price is very reasonable for a professional event. 
     AT&T is sending Melissa Arnoldi, Verizon Nicola Palmer, T-Mobile Karri Kuoppamakia and Sprint Günther Ottendorfer. Durga Malladi and Nokia’s Mike Murphy. Charter Cable is sending Craig Cowden; anyone who thinks Comcast and Charter aren't serious about 5G is uninformed. (Although buying Sprint may be their ultimate decision. News
A remarkable 400 people attended the very strong Broadband Forum BASE events in Berlin and Las Vegas. Trevor confirmed BT would pass the million this year. Cioffi projected “Waveguide DSL” could carry 10 gigabits a kilometer as well as a terabit 100 meters. Werner sees a 4X improvement in upstream with cDTA. Much more in next issue.

Deutsche Wants a Gigabit, Finally Realizes 50 Meg Isn't Enough
Deutsche Telekom is finally realizing that 50 megabit DSL won't make it against gigabit cable. VP Franz Seiser is blunt. "We must change radically, become disruptive and, above all, throw away things," he proclaims at BBWF. After years of DT insisting 50 megabits is plenty, we now hear "it is about Gigabit products" from DT's Robert Soukup.  
    A lucky building in Frankfurt will receive 500+ megabit service as ultra-conservative Deutsche Telekom experiments with Soukup told BBWF, "We're going to have a field test in Frankfurt with and Fiber To The Building (FTTB.) We will know by the end of the year if this is the right way to go." Hint to Soukup: Yes it is. is working well at a dozen telcos I;ve talked to.
     The details are surprising. DT is going for CORD, Open Source, Calix, and Radisys.

*** The new Telebyte Guide to Testing Gfast follows the Broadband Forum IR-337 Gfast test specification, the same used by the University of New Hampshire (UNH-IOL) for Gfast certification testing. Free download (ad) It is the best technical guide to  I have seen. Grab it. Dave

1.6 Gig in Sckipio-Calix Test
A telco tells me they are getting impressive early results from the Calix 48 port DSLAM with the new Sckipio 212 MHz chips. There still is work to do but this is encouraging. 
    Carriers want DSLAMs with more than 16 ports to reduce the deployment costs from the basement or larger field cabinets. Speed matters to the marketing side of the company; AT&T's CEO believes he must offer a true gigabit to match cable. (They've been getting ~750 megabits with first generation chips.

*** Self-Healing Wi-Fi With ASSIA Real-Q 
Beyond-the-Box visibility and control extends quality-of-experience (QoE) beyond the gateway to the end-user device for every device in the home. Based on ASSIA technology, proven across 80 million subscribers (ad)

Reverse Power 4 Port DSLAM for Australia
Australia is connecting 1M homes to, some with a Netcomm distribution point mini-DSLAM. It's a small unit designed for pole or pit mounting. It's waterproof, pressure proof, and temperature resistant. Their matching home modem is bittorrent friendly, with two USB ports for a hard drive dedicated to sharing.
     A reverse power unit at the customer, the NDD-0100-01, can save the cost of bringing power to the DSLAM. They don't expect many orders until the second half of 2018, as nbn is waiting for the second generation chips. Netcomm demonstrated RP with BT Openreach in August.

*** Sckipio's Three advances are taking to the next level. (ad)

Australia Makes it Official: to Million Plus
No news here. In September, 2015, I reported Australia's nbn Going This June. I reported the million home fiber to the curb (kerb?) was beginning. Unfortunately, they are no closer to figuring out where to find the needed $10B to $20B to cover the cost overruns. Instead, the parties are battling in Parliament about who is to blame.

2 Bonded 212 Lines = 3 Gigabits
Sckipio at BBWF is demonstrating 3 gigabits down, nearly a gigabit up, over two phone lines, bonded. Twice the bandwidth (212 MHz instead of 106 MHz) times two lines is fast. Sckipio does great demos; at CES, they showed first generation chips delivering almost 1 gig upstream.
    “Sckipio is pushing Gfast to astonishing speeds with production silicon,” CEO David Baum proclaims. Calix is using the SCK23000 chipset in their 48 port gig+ DSLAM at the show.


  • Carl Russo of Calix was very optimistic on the financial call, reporting increased sales every quarter. The shift of the company to software, especially SDN, is moving well. The losses on deployment services are turning around. In particular, they are more carefully defining what is included in their contracts, thereby controlling costs. That said, losses will continue through Q4. I believe increased Cloud and SDN software sales bode well going forward. Emerging customers Verizon & Deutsche Telekom have great potential.
  • “The silicon for 212 is very immature,” says Steve Collins of Netcomm Wireless, who doesn’t expect it to be ready for a year. He’s working with, and presumably discussing, the Broadcom chip.
  • The Broadband Forum will have their next BASE (Broadband Access Summit Event) in Greece, Bernd Hesse writes. The date isn’t set. 

*Reply "subscribe" to be added, "un" to be dropped
Volume 18, #7 Nov 6, 2017Oct 23 20 Gigabit DSL - Live Demo; Calix Showing 48 Port DSLAM; Will BT's 10M be Fast Enough to Protect the Company?; 212 MHz Likely Late 2018

un Zhang and his Chinese telecom colleagues are heroes. They have connected directly to fiber more than 272M homes, twice as much as the entire rest of the world. LTE coverage is 98% and nearly a billion have signed on. Prices are relatively low. 

We all know the nature of the Chinese government but I can also admire the achievements.

A practical problem has kept me in New York rather than the events in Berlin. Please keep your ears open for mail and send it to me, This email address is being protected from spambots. You need JavaScript enabled to view it..

AT&T's Possible Dream: "We Will Spend $5B Building Two of the World's Best Networks After Harvey &  Sonoma."
Randall Stephenson, AT&T CEO, did not proclaim, AT&T will spend $5B building one of the world s best networks for Texas and Louisiana. That s our contribution to the recovery after Harvey. He should just do it, now adding North California.
Larry Babbio of Verizon responded that way to the World Trade Center. "Just watch us," he replied to my suggestion Verizon bring fiber to lower Manhattan soon after 9/11. Verizon built Fios.
The fix is reportedly in for the AT&T takeover despite Trump promising, "AT&T is buying Time Warner and thus CNN a deal we will not approve in my administration because it's too much concentration of power in the hands of too few."
A good time for Randall to be a hero. much more

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." link (ad)

Hundreds of Thousands Go Wireless Only in Tokyo
Softbank was startled by large signups for fixed wireless in Tokyo, a city with almost 100% coverage of fiber home. T-Mobile in Austria has also seen a trend.
The new customers were primarily young. They wanted a service they could take with them if they changed apartments. The fiber connects and reconnects are a headache.
Softbank has nominal speeds of hundreds of megabits and practical speeds of at least a couple of dozen megabits in most of Tokyo. That is enough for most people, most of the time. more

*** Make sure to visit Adtran Booth E102 to see important new products. Please join us on Tues 10:05 Werner Heinrich DTA 10:35 Ryan McCowan NGPON2 11:20 Kurt Raaflaub Fiber to the DP without fiber 11:35 Robert Conger NGPON2; Wed 10:30 Ronan Kelly Keynote 14:00 Robert Conger Open Networking 14:20 Gary Bolton Platform Economy 16:50 Heinrich, Chris Thompson, Jeremy Harris (separate panels); Thurs 16:50 McCowen (ad)

Is Google Go-Long the Future of Fiber Home?
Cedric Lam on Oct 26 will unveil Go-Long. Google thinks "this new standard can reduce the network operating costs and investment costs." The only information is the press release,
Go-Long is the company's new generation of optical access networks. Described by Google as a new Time and Wavelength Division Multiplexed Super PON (TWDM Super PON), the company believes that this new standard can reduce the network operating costs and investment costs. The company claims that the technology can do this by extending the transmission distance, which can reduce the number of central offices that need to deploy the optical line termination (OLT), significantly reducing equipment, space and power costs.
In other words, they ain't saying anything until the announcement at BBWF.

*** Telefonica Deploys ASSIA�s GPON Expresse Solution
ClearView support for Wi-Fi and Expresse are now available for both DSL & GPON (ad)

50G & 100G Ethernet for Backhaul
Huawei has delivered a testable 50G unit to China Mobile and passed testing at the European Advanced Networking Test Center. They also offer a 100 gigabit unit with four 25 gigabit devices. A 2 x 50G unit is planned that will be air-cooled.
Verizon has chosen 40 gigabit NG-PON2 for small cell backhaul, but most other carriers are choosing ten gigabits. Verizon has been testing NGPON2 from Calix and Adtran for a year and is now deploying hundreds of 5G small cells. more

*** Columbia CITI Nov 17 State of Telecom
Eli Noam and Bob Atkinson have the very best in world telecom to the university for this annual event. It was unmissable for anyone who cared about policy. This year, it's going online, (psa)

272M Fiber Connections in China: The Most Remarkable Achievement in My 20 Years of reporting
Jun Zhang, a senior China Telecom engineer, was surprised when this New York reporter called him and his colleagues heroes. They have connected directly to fiber more than 272M homes, twice as much as the entire rest of the world.
There are well over 300M homes passed. 92% of urban homes are covered by fiber. Cities like Shanghai are close to 100% covered. Nearly all the fiber build was done in five years. Prices are relatively low. I am not blind to the authoritarian Chinese state, but I can also admire the achievements. more

Steve Crocker Exiting ICANN Chair
Steve writes, "We have term limits for directors, no more than three consecutive three year terms.  I have reached my limit and am leaving the Board." He has been active literally since the beginning of the Internet. He was part of the UCLA team, led by Len Kleinrock, that connected the first Internet node, back in DARPA days. Steve went on to create the RFC system.
The people who built the Internet are still very active, but we're all getting on in years. Steve, Vint Cerf, Dave Clark, and Bob Kahn are all in their seventies. All of them are justly proud and hopeful the Internet will continue to thrive.
Two-thirds of Internet users are not American or European; China is now one-third of the Internet. No one from mainland China has served on the boards of either ICANN or ISOC. ICANN's new board members are both American.
It is time for a Nixon goes to China moment or the old systems are unsustainable. The Chinese and the other BRICS are already building new alternatives more. News
The best in the business will be at BBWF in Berlin, many speaking at the Broadband Forum event. A practical problem leaves me back in New York, so please listen closely and send me the news.

20 Gigabit DSL - Live Demo
The meter reads 20 gigabits.  Jack Zhu had just said, "You have to see this," and pulled me over. It was the first demo of Huawei's four port bonded G.mgfast.
Two boxes perhaps five feet apart, were connected by four lines of cat 6 cable. Each line carried 5 gigabits of prestandard G.mgfast.  The total is 20 gigs, They expect that speed to go 30 meters. They are finding good performance out to 100 meters. more and a picture

*** At Broadband World Forum Berlin, make sure to visit Sckipio in Booth E101B for the latest advances in Don't miss our CTO, Rami Verbin, at the Broadband Forum Access Summit. Sckipio, the leader as deploys around the world, continues to deliver performance. (ad)

Calix Showing 48 Port DSLAM
Testing in a major telco lab. The new Sckipio chip allows daisy-chaining 24 port boards for up to 96 ports. (Other approaches require an external vectoring controller.) Calix has wasted no time developing a 48 port DSLAM, showing this week at BBWF.
The new Calix/Sckipio DSLAM also supports "up to 2 gigabits," although it best to think of that as 1.5 gigabits. The cDTA allows near instantaneous switching of an individual subscriber's bandwidth from downstream to upstream, allowing much faster upstream. more

*** Two remarkable Broadband Forum Access Summits will feature speakers including Tom  Starr of AT&T and Trevor Linney of BT. In Berlin at BBWF October 24 In Las Vegas before the Calix Connection event October 29 (psa) See you there

Will BT's 10M be Fast Enough to Protect the Company?
BT is close to a million homes passed with Trevor Linney is now leading a rapid rollout, on track for 10M in 2020. BT remains at the forefront of research and is showing impressive technical capability.
At the management level, BT's beleaguered boss Gavin Patterson quietly made a decision that is enormously risky. To save money, he canceled the original plans for a gigabit network with fiber to the distribution point. Instead, BT is putting DSLAMs at the existing cabinets, reducing speeds to 160 or 330 megabits. That's G.halffast.
New chairman Jan du Plessis may sack Gavin, more gossip

*** The new Telebyte Guide to Testing Gfast follows the Broadband Forum IR-337 Gfast test specification, the same used by the University of New Hampshire (UNH-IOL) for Gfast certification testing. Free download (ad) It is the best technical guide to  I have seen. Grab it. Dave

212 MHz Likely Late 2018
Ryan Ding of Huawei is waiting for updated Broadcom firmware, which he expects before the end of the year. He will then be able to supply units for telco trials "in the first half of 2018."
Ding and everyone else are working diligently to get the DSLAMs ready to deploy in volume as soon as possible, I think late 2018 looks most likely. We won't know for sure until we have several months of tests.
Preliminary reports are that both the Sckipio and Broadcom chips are reaching 1.5 gig. more


  • �MediaTek�s recovery will take time as its modem technology is still generations behind market leader Qualcomm.� Stuart Robinson, Strategy Analytics The 835 is pulling Qualcomm far ahead.
  • Moore�s Law, by Arnold Thackray, is an authorized biography full of details fascinating to anyone interested in chips. Gordon Moore was there at the beginning, developing diffusion at Shockley, then bringing it into production at Fairchild, supervising the birth of the planar transistor, and leading the industry for so many years. It�s written with Moore�s spirit of frankness and honesty, as well as his deep collaboration. Most of the book is about the technology and industry and is very strong.


Oct 16

CTO Seizo Inoe of NTT last year thought 2022 for mobile 5G. Now it’s 2019 for Verizon’s large network. and Qualcomm’s phone chips. The tech is moving faster than believed by anyone except the “father of millimeter wave,” Ted Rappaport.

  • I’ve built a 5G mmWave map, starting with Verizon and China Mobile. Korea, Japan, and AT&T may be next. Hundreds are in trials but most of them are pr or bogus. Almost all Europeans think the investment won’t pay off. They will do very little without government orders. 'Incentives' are bs. They rarely work.
  • Randall Stephenson should say "AT&T Will Spend $5B Building Two of the World's Best Networks After Harvey & Sonoma." Gov will pay most of it.
  • Netheads had to fight Bellheads for the right to exist 30 years ago. Subtly, the incumbents are striking back. The otherwise fine NGMN roadmap implies the telcos run everything and collect a mileage tax on every car.

Hangzhou tomorrow. Berlin and Las Vegas the next. Say hello to the round fellow with a beard, wearing a black shirt to stand out. 

Huawei and Adtran are paying my expenses. Thank you. News: 10 Gigabits 500 Meters Over Phone Cables?! ; Who needs fiber? Backhaul By a Dozen Bonded DSLs ; Bonded 35b Tester From intec ; What Will Broadcom, Huawei, Metanoia & Nokia Bring To Berlin Broadband; Newsbreak: Nokia's Hidden New SX-16F DSLAM

2019! "You’ll see 5G in 2019 for sure" Qualcomm
2019 Mobile 5G Phones! blazes on ads across the trade media from Qualcomm. "You’ll see 5G in 2019 for sure," Qualcomm CEO Steve Mollenkopf promises. Verizon has begun a huge, rapid deployment of mmWave, with hundreds of sites installed. 

Mollenkopf himself thought it would be 2020, but new technology is allowing him to chop a year off the schedule. His main foundry, TSMC, has begun "risk production" of 7 nm chips, the next generation. It's reasonable to expect volume production in 2019. TSMC is taking delivery in 2018 on $150M EUV machines to improve critical layers. Steve is going for 2019 because he can.

Verizon will be ready to serve mobile 5G "as soon as phones are available." China Mobile and AT&T are also looking to build early. They will be expensive and probably power hogs, but a true gigabit should be possible. ("Gig LTE" will usually be 75-300 megabits to the customer, per Qualcomm.)

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." link (ad) 

Algeria Fibering 1M. Fiber Ain't Dead
AT&T is fibering 3M homes/year. Telefonica doing millions in Brazil and most of Spain. France Telecom is well along and nearly all of Portugal covered. Bell Canada recently decided to speed up fiber because cable was killing them. 

After a year of political wrangling, state-owned Algérie Telecom has signed a contract with Huawei and is ready to go. The initial contract is 1M homes servable, with a plan to go to 3M. Wireless networks spread incredibly rapidly across Africa and now fiber (mostly backbone) is also building.

*** Make sure to visit Adtran Booth E102 to see important new products. Please join us on Tues 10:05 Werner Heinrich DTA 10:35 Ryan McCowan NGPON2 11:20 Kurt Raaflaub Fiber to the DP without fiber 11:35 Robert Conger NGPON2; Wed 10:30 Ronan Kelly Keynote 14:00 Robert Conger Open Networking 14:20 Gary Bolton Platform Economy 16:50 Heinrich, Chris Thompson, Jeremy Harris (separate panels); Thurs 16:50 McCowen (ad) 

NGMN Has Defined a Wireless Future - by the Telcos
20 top telcos banded together to set a common path for the next thirty years. China Mobile, Deutsche Telekom, AT&T and many others have now produced the NGMN 5G End-to-End Architecture Framework O.8.1. The 36 pages are a blueprint of how the telcos want the future. They've done an extraordinary job making the concepts clear to anyone knowledgeable in today's wireless. 

This is the antithesis of the Internet model, a network of networks connected but operating independently. Implicitly, everything would be under the control of the telcos. They would have standard procedures to work together for international traffic, based on everyone having common QoS control.

Almost no one is noticing. The Bellheads are striking back.

*** CloudCheck® Wi-Fi optimization platform enables visibility, control, and self-healing of residential Wi-Fi networks. Our new Mobile APP for Service Providers  provides a set of powerful and easy-to-use performance and install assist diagnostics tools for field (ad)

Why are 3 Indian Survivors Spending $B's on Massive MIMO?
Competition is fierce and spectrum is short, Reliance Jio spent $20B, built one of the best networks in the world, and gained 100M customers in the first year. Mukesh Ambani's aggressive plans for the 4G $20 (refundable) feature phone will probably add 100M more in the next year or so. Hundreds of millions of Indians are still on 2G. Ambani is ready to spend $B+ in subsidies to bring them to Jio 4G. 

Giant Tata is closing their #4 mobile network, putting 10,000 out of work, writing off billions, and selling off cheaply to Bharti.. #5 Reliance Communications, Ambani's brother's company, is likely to follow. They couldn't compete in the price war. Vodafone and Idea, #2 & #3, are merging to stay alive  

#1 Bharti's strategy in response includes a huge build of Massive MIMO. The other two likely survivors, Jio and Vodafone/Idea, will also do Massive. much more

*** At Broadband World Forum Berlin, make sure to visit Sckipio in Booth E101B for the latest advances in Don't miss our CTO, Rami Verbin, at the Broadband Forum Access Summit. Sckipio, the leader as deploys around the world, continues to deliver performance. (ad)

5G Maps
5G is here, Verizon has installed "hundreds": of mmWave cells, per CEO Lowell McAdam. I am  trusting that China Mobile is also starting. Time to create a 5G mmWave map. 

Massive MIMO is expanding rapidly. Softbank in Japan and China Mobile are installing thousands of towers. I predict huge growth in M MIMO in 2018 & 2019. Korea Telecom is building rapidly a system along Highway 50, ready to astound the world at the Olympics in February. 

I don't include “tests” or "trials," of which there are hundreds. Too much is pr. Countries only go on these maps when I'm confident they are building and ready to serve customers. Updates and improvements very welcome:

Stories not here:
Why Verizon is rushing to 5G mmWave ** Qualcomm makes  great chips but Taiwan is right on unreasonable royalties. ** AT&T wants > 10 ms. 1 ms is likely far away. ** EU moving ahead on 25 year licenses as monopoly spectrum is becoming obsolete. Dumb. ** India’s fiber to 250,000  villages is growing. ** Totally wasteful $3B subsidy for AT&T LTE ** $20B fabs from TSMC and Samsung protect Moore’s Law to ~2025

*** Columbia CITI Nov 17 State of Telecom
Eli Noam and Bob Atkinson have ther very best in world telecom to the university for this annual event. It was unmissable for anyone who cared about policy. This year, it's going online, (psa) News
Some of the best in the industry will be at the Broadband Forum’s BASE in Berlin and Germany and BBWF in Berlin. See you there.

10 Gigabits 500 Meters Over Phone Cables?!
John Cioffi theorizes that 1 Terabit/s over 100m is possible. Several world-class engineers have confirmed the theory appears correct. The same higher-order waveguide modes could deliver 10 Gbps at 500 meters, which is exciting some of the most important companies in the industry. It could drastically reduce the cost of backhauling 5G small cells in the required millions of locations. 

Some are using the name Surface Wave over Copper (SWoC) instead of TDSL as the system moves closer to test. Building a test system and proving the real challenges can be met will be expensive. The frequencies involved are at 50 GHz and higher, 200 times higher than the new 212 MHz

*** At Broadband World Forum Berlin, make sure to visit Sckipio in Booth E101B for the latest advances in Don't miss our CTO, Rami Verbin, at the Broadband Forum Access Summit. Sckipio, the leader as deploys around the world, continues to deliver performance. (ad)

Who needs fiber? Backhaul By a Dozen Bonded DSLs
12 bonded lines of 35b VDSL should be able to deliver 2 gig downstream ~400 meters. That's comparable to the most common backhaul, 2.4 gig GPON. Telcos are confident few will use the high speeds simultaneously so the high over-subscription is realistic. Upstream on the older DSLs is lower, so the upstream will probably be limited.

Kurt Raaflaub says Adtran has trial units at customers. As far as I know, this is the first public description of bonded 35b and I hadn't anticipated this possible use. Deutsche Telekom says 35b will be ready for deployment the second half of 2018.

*** Two remarkable Broadband Forum Access Summits will feature speakers including Tom  Starr of AT&T and Trevor Linney of BT. In Berlin at BBWF October 24 In Las Vegas before the Calix Connection event October 29 (psa) See you there

Bonded 35b Tester From intec
intec and Adtran have working parts for testing bonding. 35b uses double the spectrum for double the speed, 200+ megabits at 500 meters. It's not ready for the field yet, but test gear is now available from the German firm intec.  Look for them at the Broadband Forum Interop exhibit.

Jeff Waldhuter of Verizon told me not investing in testing was the biggest mistake they made in the early days of DSL. BBWF will also feature Lincoln Lavoie, who has led the UNH/Broadband Forum testing for years. You can now buy the testing software they use for your own work. UNH uses Telebyte hardware, who will speak at the show.

*** Make sure to visit Adtran Booth E102 to see important new products. Please join us on Tues 10:05 Werner Heinrich DTA 10:35 Ryan McCowan NGPON2 11:20 Kurt Raaflaub Fiber to the DP without fiber 11:35 Robert Conger NGPON2; Wed 10:30 Ronan Kelly Keynote 14:00 Robert Conger Open Networking 14:20 Gary Bolton Platform Economy 16:50 Heinrich, Chris Thompson, Jeremy Harris (separate panels); Thurs 16:50 McCowen (ad) 

What Will Broadcom, Huawei, Metanoia & Nokia Bring To Berlin Broadband
Companies time releases for BBWF, the biggest event of the year. Here are some possibilities for the companies that haven't provided me news yet.

Huawei will brief me at their Hangzhou event this week. For now, I know the 96 port external vectoring boxes are doing well in customer labs and probably ready. Two telcos were impressed.

(Huawei, Calix, and Adtran are very generous with the press, hosting regular events. All three do a good job trying to answer questions when they can. The result is they get far more coverage. I make a point of reaching out to those who tell me less, including writing this as a reminder.) 

Metanoia is the third chipmaker that joins the UNH interops. Nokia and Broadcom have yet to speak so I make a few guesses.

*** The new Telebyte Guide to Testing Gfast follows the Broadband Forum IR-337 Gfast test specification, the same used by the University of New Hampshire (UNH-IOL) for Gfast certification testing. Free download (ad) It is the best technical guide to  I have seen. Grab it. Dave

Newsbreak: Nokia's Hidden New SX-16F DSLAM
Buried deep in an SDAN announcement is a mention of a Lightspan SX-16F. They call it  "The world's first 16-port reverse-powered micro-node which can be safely reverse-powered from the home." No more details available at press time.

*** Two remarkable Broadband Forum Access Summits will feature speakers including Tom  Starr of AT&T and Trevor Linney of BT. In Berlin at BBWF October 24 Las Vegas before the Calix Connection event October 29 (psa) See you there

Broadband Forum BASE: Berlin Oct 24 Las Vegas October 29
Free BASE? Yes, no charge to hear 30 of the most informed people in the industry/  Trevor Linney of BT, who tells me they are pas 550,000 homes, a million are close, and the build to ten million is going well. David Titus and Tom Starr of AT&T, moving aggressively to buildings across the U.S.Vincent O'Bryne of Verizon, the leading proponent of NG-PON2. 

See you there. Look for the round fellow with a beard and point me to news I missed.

*Reply "subscribe" to be added, "un" to be dropped
Volume 18, #5 Oct 16, 2017

Oct 8 Randall's Xmas: 144 Gigabits, 212 MHz, 96 Port, DTA; More Gear Passes Interop at UNH; More Gear Passes Interop at UNH Broadband Forum BASE Berlin Oct 24, Las Vegas Oct 29;

Reply "subscribe" to be added, "un" to be dropped

This is not a test.

Lowell McAdam of Verizon has "100's of mmWave cell sites up and operating," he told us at Goldman Sachs September 13, 2017. This is the true dawn of 5G mmWave. I expect thousands more in 2018 and tens of thousands in 2019-20.

AT&T CEO Randall Stephenson promised America, �A ubiquitous gigabit in the next few years.� �Gig LTE� is puffery but even 100-300 megabits everywhere is historic.

India�s Sunil Mittal has turned on Massive MIMO at Bharti. Test speeds on Massive MIMO are almost a gigabit on 20 MHz (shared.) Ericsson�s Nishant Batra expects 100 MHz systems in 2018, with likely peak speeds well into the gigabits. Three other Indian giants are also going Massive

Imagine what the Internet will be like when tens of millions have a true gigabit mobile. Lowell believes, �It's on our doorstep.�

How will we bring this to (almost) everyone worldwide?

Say hello at Huawei Hongzhou next week, then Berlin Broadband Forum and BBWF, and the Forum�s Access Summit in Las Vegas. I�m a round fellow with a beard. News, below, features AT&T�s enthusiasm for the new 1.5 Gigabit chips and reverse power.

Verizon: "100's of mmWave Cells Installed"
September 13's announcement by Lowell McAdam will go down in history as the date of the first real 5G mmWave deployment. True speeds of a gigabit are possible, rather than the 100-300 megabits more common with "Gig LTE."

"We literally have hundreds of cell sites that we have up and operating and 8 markets moving to 11 markets. ... this platform is not two or three years out the way it was a year ago, it's on our doorstep and I think it's going to be huge."

Verizon needs 5G more than any other company on Earth. Its business model is based on being better and able to charge a premium price. AT&T and especially T-Mobile have caught up in important ways.  More

*** "I'm very proud of our new Amendment 3 chipset," says David Baum, CEO, Sckipio. "The performance is breathtaking, Compared to our previous chipset, It delivers four times the vectoring capability, twice the bonding speeds, and supports coordinated DTA for symmetric-like performance in both downstream and upstream." link (ad)

95% 4G LTE For Myanmar? Yes!
MYTEL, the new fourth company, has promised 4G to 95% within three years, investing well over $1B. They will launch their 4G-only network with 7K base stations across most of the country early in 2018.  The wireless explosion in Myanmar is almost miraculous. Six years ago, fewer than 1% had mobile phones. Today, Bloomberg estimates 90% are connected.

90% to 95% LTE coverage is rapidly becoming common across the developing world. Reliance Jio will cover 95% of India soon. The U.S. officially is over 99% LTE, although some think the figure is high.

Government owned networks are increasingly important. Viettel will own 49% of MYTEL. The remaining 51% will be split between MEC and local companies. MEC is controlled by the Myanmar military and Viettel by the Vietnamese government. Today, ~3/4ths of new wired broadband connections are on government-controlled networks, led by the Chinese. More, including a comment on the Myanmar military,

*** CloudCheck� Wi-Fi optimization platform enables visibility, control, and self-healing of residential Wi-Fi networks. Our new Mobile APP for Service Providers  provides a set of powerful and easy-to-use performance and install assist diagnostics tools for field techs. (ad)

Historic Promise: AT&T Soon Will Offer "1 Gig Speeds Ubiquitously." Mostly True
Randall Stephenson, AT&T CEO, told us at Goldman date "Over the next few years you get 1 gig speeds ubiquitously." AT&T's actual plan doesn't go that far but is remarkable. Some percentage of the country - ?5-15% - will not be covered. The "gig" will be 75 to 300 megabits down, most times and most places. That's not too shabby by any standard and a huge jump over the 5 meg to 25 meg speeds most common on U.S. LTE networks. Randall sees truth like a politician.

In wireless, I assume he mostly is planning "Gig LTE." In certain test conditions, that does deliver close to a gig. In the real world, the best analysis available indicates speeds will usually be a quarter gig or less. (Article coming.) The other three six big U.S. and Canadian carriers have similar plans. So do major carriers in Myanmar, India, and many other places.

On the wired side, AT&T is adding about 3M homes passed with GPON each year, which delivers a true gigabit down and 500+ meg up. Within about five years, 30-50% iof AT&T homes will be able to get 800+ meg downstream, close enough to a gig for most practical purposes. That will include 15M+ with fiber home.  More

*** Two remarkable Broadband Forum Access Summits will feature speakers including Tom  Starr of AT&T and Trevor Linney of BT. In Berlin at BBWF October 24 In Las Vegas before the Calix Connection event October 29 (psa) See you there

99.5% U.S. LTE? 98%? 96%
Senator Schumer wants $40B to the many millions he thinks can't get broadband. Pai has $4B in the pipeline for areas without 4G. Totally political and likely to waste $billions. LTE is at 10-25 meg today and AT&T promises a "ubiquitous gigabit." Coverage isn't perfect but there aren't that many populated gaps.

The growth of LTE networks continues to astonish even those of us who meekly predicted near-ubiquity as far back as 2009, US Telecom's Pat Brogan estimates LTE coverage at 99.5% of the population, based on company data reported to the FCC. Both Verizon and AT&T claim 98% coverage and T-Mobile almost as high. They have different unserved areas, so 99.5% amongst them is plausible.

The real figure may be as low as 95-96% because coverage data is notoriously unreliable, with companies having incentives to inflate claims. I remember AT&T telling the FCC a while back the FCC didn;t need to test wireless availability because the companies could provide accurate maps. I checked AT&T's map of Manhattan and saw it claimed as 100% covered. I live here, and at the time was hearing numerous complaints of deadspots.

*** Columbia CITI Nov 17 State of Telecom
Eli Noam and Bob Atkinson have ther very best in world telecom to the university for this annual event. It was unmissable for anyone who cared about policy. This year, it's going online, (psa)

Ericsson: FDD Massive MIMO Ready, Not Just TDD
Time Division Duplexing has dominated Massive MIMO, including the thousands of cells at China Mobile and Softbank. Most telco spectrum is dedicated to Frequency Division Duplexing.

Massive MIMO inventor Tom Marzetta was skeptical about the performance of Massive MIMO using FDD (Frequency Division Duplex) rather than TDD (Time Division.) Massive MIMO requires regular and robust reporting from the cell phone, which they feared would be impractical with FDD.

Huawei, ZTE, and now Ericsson have done field trials they believe prove otherwise.  In trial with T-Mobile. Nishant Batra is confident many antenna Massive MIMO will add to performance in FDD frequencies as well as the TDD frequencies beginning to widely deploy.  I'm waiting for performance data, however.

Batra expects 3 carrier, 60 MHz TDD Massive MIMO in early 2018. Batra expects 100 MHz Massive MIMO by yearend 2018.

==================================== News - AT&T Sees 1.5 Gigabits
Eric Small of AT&T tells Sean Buckley �In generation 2, you get about 1.5 Gbps of throughput,� confirming what the vendors promise. �Not that you would use more than a gig in either direction, but it gives you the flexibility to use a gig in one direction and 500 Mbps in the other direction.�

He adds, ��The really interesting part of the second generation of Gfast is it allows for reverse powering, which means the power could come from the individual subscribers.� Fierce Telecom

Randall's Xmas: 144 Gigabits, 212 MHz, 96 Port, DTA From Sckipio
Randall Stephenson at AT&T "needs 1 gig speeds."and promises "1 gig speeds ubiquitously." Sckipio is ready with 212 MHz Amendment 3 chips soon to sample to selected customers - including AT&T suppliers. Using 212 MHz of spectrum almost doubles the capacity of, reaching close to 2 gigabits between upstream and downstream in lab conditions.

96 ports at 1.5 gigabits is 144 gigabits. It should be achievable over the short loops typical in apartment buildings, although they will nearly actually all be a maximum. Adtran and Huawei have demo'd external boxes that can vector 48-96 ports, but Sckipio does not require the complexity of external vectoring. Instead, the first DSLAM supports 24 or 48 ports and can be daisy-chained with others on demand. This reduces the initial capex. cDTA allows switching between upstream and downstream for each line. Much more

*** The new Telebyte Guide to Testing Gfast follows the Broadband Forum IR-337 Gfast test specification, the same used by the University of New Hampshire (UNH-IOL) for Gfast certification testing. Free download (ad) It is the best technical guide to  I have seen. Grab it. Dave

More Gear Passes Interop at UNH
Lincoln Lavoie sees "cross chipset interoperability and performance at all levels including device software for management and control." Gear with chips from Broadcom, Metanoia, and Sckipio are working with each other at a good speed, usually 600 megabits or higher. EXFO and Viavi testers also passed.

He adds, "Gfast testing has already been more rigorous than any previous certification testing and we are testing individual features more deeply than we have on any previous technology."Interoperability" is considered to mean working reasonably together, not necessarily at maximum performance levels.

*** At Broadband World Forum Berlin, make sure to visit Sckipio in Booth E101B for the latest advances in Don't miss our CTO, Rami Verbin, at the Broadband Forum Access Summit. Sckipio, the leader as deploys around the world, continues to deliver performance. (ad)

Tom Starr, Trevor Linney Headline Berlin Oct 24
Linney is deploying past ten million British Telecom homes, with more connecting daily. Trevor, speaking for BT, has always been open about the practical results. Starr has led DSL standards and the Forum for 25 years, seen it all and done it all. AT&T is moving aggressively on from basements around the U.S., so he too will have lessons from the field.

The new Broadband Forum Access Summit is free. Register through the BBWF website at BBWF registration on site wasn't free last I looked so make sure to register in advance. The Forum is doing a sister event in Las Vegas on Sunday, October 29th. See you at both. Look for the round fellow with a beard. After Robin Mersh opens the event, Roland Montagne of iDATE, a respected analyst, will review deployments around the world. See you there

Broadband Forum BASE Las Vegas October 29
Reverse power. DTA. Profile 212. Now is the time to start taking advantage of the advanced features, as deployments are approaching one million homes passed. Registration is free at See you there. Look for the round fellow with a beard. David Titus of AT&T will present details from their nationwide effort.

*** Self-Healing Wi-Fi With ASSIA Real-Q
Beyond-the-Box visibility and control extends quality-of-experience (QoE) beyond the gateway to the end-user device for every device in the home. Based on ASSIA technology, proven across 80 million subscribers (ad)

Scared By Cable, Bell Canada Chooses Fiber Over
Better TV software long kept Bell's DSL ahead of cable, but both Shaw and Rogers are now deploying Comcast's respected X1 software and will likely catch up. CEO Cope hinted to Wall Street he has little choice but to find the necessary capital. was considered until very recently but Bell decided to accelerate their fiber home. They have done  most of Montreal and are working on Toronto.

Reply "subscribe" to be added, "un" to be dropped
Volume 18, #4 Oct 8, 2017

August 27

Reply "subscribe" to be added, "un" to be dropped AT&T: Millions of Lines Coming. Now Starting Buildings Outside of Territory; Here Comes 5 & 10 Gigabit G.mgfast; 424 and 848 MHz ; Stanton of Adtran: Will Be Even Faster, But Not Yet ; Broadcom's 212 MHz Exists Update: I'll have test data next issue; 1.6 Gig 212 MHz ZTE/NetCologne Demo; Broadcom Bummer: Blocked By Berlin Ban; Tamboli: "2019 Will Be The Year of"; 2019 Deutsche Telekom Build is On below

25,000 jobs are going at Ericsson, a tragedy. 70,000 cuts are in process at AT&T. Please don’t shoot me for the message. Blame the engineers who are improving gear faster than demand is growing. 

AT&T with 140,000 employees rather than 210,000 still will be a giant, even if profits are only $10B instead of $20B. Let all live long and prosper but it’s time to accept that telecom is a shrinking industry. China Telecom is still growing but even they are cutting capex by 8%. 

Low prices are great for consumers but brutal for the people involved. Rapidly improved technology often devastates workers. Most Thai farmers suffered when miracle rice doubled production and English textile workers suffered in the time of great satanic mills. 

A special policy issue will begin: Microsoft President Brad Smith was wrong claiming "Today 23.4 million Americans in rural areas still lack broadband internet access." The figures are from 2014. Probably half of them are now connected with LTE ...  Or should I say "Brad Smith lied?" Why does Germany have some of the slowest connections in Europe? How much of the $8B spent on U.S. rural subsidies was wasted? Has India found a way to connect 300M villagers? I’ll do my best to get close to the truth. 

The Samsung 8 and a half dozen other phones with the Qualcomm 835 support Gig LTE. It’s really “Quarter Gig,” with likely speeds 100-300 megabits. That’s still darn good. More than two dozen telcos will hype Gig LTE networks soon, including all three in Korea and all four in the U.S. 

Huawei is close to a prototype of a 2 Gig LTE phone, using 5 20 MHz bands compared to the 3 bands common today. 

AT&T's project is moving rapidly, although Verizon and Bell Canada are going with fiber home & wireless. A <well-informed source> tells me Germany is jumping in with in 2019. BT has already passed 500,000 homes, with a million due by the end of the year. Australia will do a million as well.

5G: Comcast, Liberty Most Serious About mmWave
The gleam in Rob's eye says he wants to go ahead. "We have more fiber than anyone else," Rob Howald of Comcast claims, adding, "We and all the others in cable have deployed fiber deeply throughout our networks. That will give us a crucial advantage in 5G mmWave." Howald was very clear that no decision has been made to go forward, although he was obviously enthused about the technology. It's not news that cablecos are looking, but I was amazed at how advanced the preparations are. 

Balan Nair at Liberty Global, the largest international, a few months ago was deeply skeptical of building his own network. Now, he tells WSJ,  "We have fiber to many neighborhoods." Stu Woo points out, "The company has utility cabinets in neighborhoods already connected to power, which would allow the company -- or a mobile-carrier partner -- to quickly set up a 5G cellular site there."

A few days later, I saw  Howald hold up a chart of 5G coverage simulations. Their current fiber allows mmWave to more than half the area studied and they are driving fiber deeper. I've confirmed U.S. #2, Charter, is thinking similarly. The final decision is a couple of years away, but the engineers are enthusiastic about the challenge.

*** At Broadband World Forum Berlin, make sure to visit Sckipio in Booth E101B for the latest advances in Don't miss our CTO, Rami Verbin, at the Broadband Forum Access Summit. Sckipio, the leader as deploys around the world, continues to deliver performance. (ad)

Ericsson Cutting 25K Jobs
Svenska Dagbladet report not denied. Every large telco vendor outside of China has been suffering for several years, although Ericsson maintained an illusion until recently. Ericsson and Nokia remain remarkable companies with world-class engineers, but it's a tough business. An Ericsson exec in India estimates the world market is going down currently at ~5%/year. 

Official details are limited. Ericsson claims the cuts "do not include R&D." Neither Ericsson nor Nokia can cut research. They face Huawei, with an R&D budget larger than the combined spending of Ericsson, Nokia, and whomever you place third. All three have great products, but Huawei's greater research budget is allowing them to outflank others with a wider product line. 

The service division is the largest target. It's probably been losing money for several years but that's been buried in the financials. The video division, built up by spending billions, is on the block. That includes what once was Microsoft's IPTV.

*** The new Telebyte Guide to Testing Gfast follows the Broadband Forum IR-337 Gfast test specification, the same used by the University of New Hampshire (UNH-IOL) for Gfast certification testing. Free download (ad) It is the best technical guide to  I have seen. Grab it. Dave

Q2 U.S.: Every Telco Goes Down
Cable up to 64% of the market. For a decade, telcos held on to 45% of the market in the U.S. But AT&T and Verizon haven't upgraded almost half their lines, planning to go wireless only to 10's of millions of homes. (One network is cheaper than two.) The other U.S telcos have done even less, keeping dividends far higher than earnings and skimping on capex.

Customers are fleeing DSL networks that often are ten and fifteen years old. The U.S. has cable to 92%, with most offering 50 megs standard. Comcast's primary offering is 100 megabits and they just extended a gig of DOCSIS 3.1 to Boston and Philadelphia. Frontier lost 101,000 subs and Century 77,000. Verizon's once world-leading fiber network wasn't  to prevent a loss of 23,000 customers. The saddest story is AT&T. They have added 5M lines of fiber the last two years and spent $67B on DirecTV, seeking an expanded customer base. Despite that, they lost 9,000 customers

*** Self-Healing Wi-Fi With ASSIA Real-Q 
Beyond-the-Box visibility and control extends quality-of-experience (QoE) beyond the gateway to the end-user device for every device in the home. Based on ASSIA technology, proven across 80 million subscribers (ad)

Cox 40% Gigified in 2017 (Correction)
All 6M Cox customers should have a gig available by the end of 2019.  Correction: I had reported that Cox had intended to offer a gig to all in 2016. I believed that was a commitment to a consumer priced DOCSIS 3.1 service. I was wrong. Cox did have a near-universal gigabit offering, but it was a business priced dedicated fiber, ($thousands, I believe.)  

Todd Smith of Cox writes, "DOCSIS 3.1 is live now with employees and we will begin deploying to customers in some markets later this year. 40 percent of the households we serve nationwide will have access to Gigabit speeds by the end of 2017." Cox Vice President Philip Nutsugah tells Mari Sibley, "The company plans to reach the rest of its customers (or at least 99% of them) with gigabit broadband by the end of 2019."

50M home giant Comcast expects, "DOCSIS 3.1 across the majority of its footprint by year-end."

16+ Megabit White Space Radios Reaching the Field
I believe this is the first report anywhere of more than 10 megabits to customers.  Michael Davies and Richard Yu of 6Harmonics in Ottawa have sent me test data showing customers with a connection phy rate of 16-18 megabits, as well as convincing details from other deployments doing better than that. I thank Boston Consulting Group for pushing me to go beyond published reports and get these new results. The fuzzy photo below shows eight users connected to a base station using a single 6 MHz channel. They have deployments in California and North Carolina using two channels for nearly double the speed. High speed uplink is included.

They are ready with a three channel unit for even higher speeds. Yu has been working on multiple antenna systems (MIMO) for almost 20 years. I'm sure he can achieve even better throughput using more antennas. They seem to be so busy actually building the equipment they haven't had time to get the latest results up on their web site. CEO Yu, off the record, shared remarkable predictions for what they will offer within a year.

*** Two remarkable Broadband Forum Access Summits will feature speakers including Tom  Starr of AT&T and Trevor Linney of BT. 
In Berlin at BBWF October 24agenda/1/grid
In Las Vegas before the Calix Connection event October 29 (psa) No charge

DragonWave Quenched, Alas
Huawei and Adtran are perhaps the only consistently profitable companies as overall sales continue to fall. DragonWave, a microwave pioneer, couldn't overcome industry problems. Ex-CEO Peter Allen is one of the go-to people in mobile backhaul. He built DragonWave up to 150M in sales in 2015, although profitability has long lagged. Two years later, three-quarters of those sales are gone. Creditors forced the company into receivership. Allen and the other directors resigned.

James Bagnell has a thoughtful report on what happened. Half of DragonWave's sales had come through a deal with Nokia. They disappeared when Nokia bought Alcatel and switched to Alcatel microwave gear.  Technical problems killed the company's sales in fast-growing India. Allen ran out of options for raising money for a comeback. The remarkable success of Wi-Fi backhaul, especially Ubiquiti, is reducing demand for traditional microwave gear.

Irrelevancy: If you want to reduce your risk of heart attacks, read Kathiresan and Topol on Genomics of Heart Disease Two world-class researchers speaking informally. You won’t be surprised that exercise and weight loss are important, but there was a lot here new to me. Separately, the data from 521,000 people is persuasive in Coffee Drinking and Mortality in 10 European Countries: A Multinational Cohort Study News

AT&T's decision tells the world is ready to go. Deutsche Telecom intends to join them in 2019, according to a source I trust enough to print the story. Verizon, citing a history of VDSL interop problems, will go with a little bit more fiber and especially their $20B 5G mmWave project rather than Bell Canada recently decided it would be all fiber home from here, as Shaw and Rogers cablecos become aggressive.

AT&T: Millions of Lines Coming. Now Starting Buildings Outside of Territory
T plans much bigger splash in territory starting soon as well. Large buildings and groups are being fought over hard by AT&T, Verizon, Comcast, and others. Verizon brought Fios fiber to this market almost a decade ago, as did local independents. Comcast is offering gigabit coax, which it is soon offering to all ~40M homes as well. 

AT&T jumped in more recently, led by Ed Balcerzak of their DirecTV division. DirecTV has millions of lines of coax bringing TV from rooftop antennas, a natural opportunity for broadband. Reporters see the new announcement as the kickoff of a major move with In selected neighborhoods of ~65% of the U.S., AT&T will consider running fiber to the basement or WTTR - Wireless to the Rooftop - plus at 500 megabits.

*** The new Telebyte Guide to Testing Gfast follows the Broadband Forum IR-337 Gfast test specification, the same used by the University of New Hampshire (UNH-IOL) for Gfast certification testing. Free download (ad) It is the best technical guide to  I have seen. Grab it. Dave

Here Comes 5 & 10 Gigabit G.mgfast; 424 and 848 MHz
"Multi-Gigabit Fast Access to Subscriber Terminals" is a new proposed ITU standard. 2020 is the target date for deployments. Jochen Maes of Alcatel/Nokia has been bringing his prototype 5 & 8 gigabit systems around the world since 2014, impressing telcos from Germany to Australia. He uses more frequencies, full duplex signalling, and other optimizations to get extraordinary performance. Adtran's simulations are that 4 gigabits can go ~75 meters using 424 MHz. 848 MHz can deliver 8 gigabits ~30 meters. (Phone wire, ideal conditions. Coax would have longer reach.)

Either would be great for the 55 families in my six story New York building. The simulations from Adtran and others in the standards groups have consistently predicted what the actual performance would be.

*** At Broadband World Forum Berlin, make sure to visit Sckipio in Booth E101B for the latest advances in Don't miss our CTO, Rami Verbin, at the Broadband Forum Access Summit. Sckipio, the leader as deploys around the world, continues to deliver performance. (ad)

Stanton of Adtran: Will Be Even Faster, But Not Yet
Customers want 212 MHz but development is still needed. Tom hopes for the first half of 2018, but notes, "The timetable for volume deployment is dependent on the chip suppliers." One European telco was told not to expect much until well into the year. I wouldn't be surprised if few customers are connected until late in the year. 

Adtran demonstrated two DSLAMs with Deutsche Telekom. One used Sckipio chips and delivered DTA. DTA allows higher effective speeds by allowing different upstream/downstream splits for each user. Since most systems are configured for a 5-1 or higher upstream/downstream split, the improved upstream can be 5X. The second DSLAM used Broadcom chips and 212 MHz of spectrum but without DTA.

*** ASSIA: Network software that ensures your customers get the performance they demand, (ad)

Broadcom's 212 MHz Exists Update: I'll have test data next issue
Deutsche Telekom and NetCologne have tested it. Adtran thinks they will be able to ship soon, but NetCologne doesn't expect much until next year. Broadcom's Greg Fischer confirms to me, "We entered production with 212 MHz G.Fast devices last quarter and also released production standards compliant iDTA.  We’re working with standards groups to assure same with cDTA." Sckipio is hard at work on 212 MHz and is delivering cDTA. It will be reassuring when the two demonstrate interop, as the lead at a very large telco discussed with me.

212 MHz should deliver as much as a gig up to about 200 meters.That covers most buildings. Almost all buildings would be covered by a second system deployed for the upper stories.

Broadcom has a "policy of not publishing data on our devices."  Vectoring noise cancellation is of course a challenge at the high speeds, so everyone will be waiting for demonstrations with a full binder. Australia and others are interested in remote power, also tbd. Broadcom has done innovative work with local error correction, and I'd like to see that shown as well.  

Traditional industry practice is to have a second source. Broadcom has just been locked out of the crucial German market in a patent dispute. They also lost a preliminary decision at the U.S. Trade Agency. They happen to be right on the substance; Tessera is demanding an unreasonable royalty. But that holds little weight in the absurd patent system. Stuff happens. A factory can catch fire. A huge customer can demand priority. A development program can fall behind. (35b was promised for 2016. DT now says it won't be ready until h2 2018.) Everyone wants to be protected.

1.6 Gig 212 MHz ZTE/NetCologne Demo
Next year, 260,000 lines. Cologne and Munich city nets have long been outclassing DT, especially with fiber deployment. CEO Timo von Lepel gets it, "One day, 100 Mbps will be the lowest data rate available." They have installed 25,000 kilometers of fiber, including to the basements of most substantial buildings. They run VDSL at 50-100 megabits to most apartments, but will rapidly replace that with

They plan to use 212 MHz of bandwidth for speeds over 1 gigabit. At Anga Com, they showed download speed of 1.6 gig and upload of 200 megabits with equipment from ZTE. In the picture next to the speed test are Sun Jie of ZTE on the left, von Lepal and Horst Schmitz on the right. They have good reason to smile. Apparently, ZTE is using Broadcom chips. They expect the 212 MHz chips to be in limited supply until next year.

I'm waiting for Broadcom to provide details. In particular, it's not clear Broadcom (or anyone) is ready to do the heavy calculations for vectoring the higher speeds on more than 4 ports. Separately, Adtran and Deutsche Telekom announced results of their private testing of 212 MHz and cDTA. The cDTA is on Sckipio chips. 

*** Self-Healing Wi-Fi With ASSIA Real-Q 
Beyond-the-Box visibility and control extends quality-of-experience (QoE) beyond the gateway to the end-user device for every device in the home. Based on ASSIA technology, proven across 80 million subscribers (ad)

Broadcom Bummer: Blocked By Berlin Ban
The Tessera patent case pushes dozens of Broadcom chips out of the country. U.S. could be next.  The German court upheld a Tessera patent "for smoothing the planes of semiconductor chips." The court ruling resulted in this notice on the Broadcom website: *Certain versions of this device are not available for sale in Germany or shipment to or from Germany, and these versions should not be used in any product destined for the German market."

It affected dozens of Broadcom chips. Broadcom asserted in court "It will suffer immense damages."Tessera has also won a patent decision at the U.S. ITC which could block Broadcom in the U.S. after the commission reviews it. The ban includes versions of the BCM63138, which was probably used in the recent Deutsche Telekom 212 demonstration.More, including the absurdities of the patent system,

2019 Deutsche Telekom Build is On
Officially mum as they seek government money. In the 2/3rds of Germany that can get cable, DT has been falling behind for years. Mike Fries of Liberty has aggressive plans to expand the cable footprint; LGI's expansion in England and Germany is the largest new wireline network building in Europe. City networks in Munich and Cologne are also upgrading rapidly. 

DT is carefully not discussing their plans as they negotiate with the government for billions in subsidies for vectoring. The 20% of Germany without cable or DT upgrade plans has $4B in subsidies already allocated. DT wants more money, but doesn't want to build a robust network where they have a near-monopoly.

*** Two remarkable Broadband Forum Access Summits will feature speakers including Tom  Starr of AT&T and Trevor Linney of BT. 
In Berlin at BBWF October 24
In Las Vegas before the Calix Connection event October 29 (psa) No charge

Tamboli: "2019 Will Be The Year of"
The technology available earlier but Dell'Oro analyst thinks telcos will move slowly.  Adtran, Broadcom, Huawei, Sckipio, & ZTE are demonstrating the next generation of gear, which presumably will be generally available in early 2018. AT&T, BT, Century, and others have plans to move aggressively, but Tamboli believes the large volume will be in 2019.

'Operators are holding off on massive deployments throughout their networks until they have more hands-on time with amendment 3 chipsets and systems, which will be available in early 2018," he believes. "Furthermore, many operators that wish to deploy into larger buildings via FTTB architectures are waiting for 32 or larger port units"

*** Virtual fiber by Sckipio. Extend your fiber with 100-300 meters of single-port It can save expensive trenching for cell towers, small cells, basement fiber, commercial customers and others. A very thin management layer allows operators to keep their existing GPON management layer. Sckipio makes it effortless to add to any GPON network.

Reply "subscribe" to be added, "un" to be dropped
Volume 18, #3 July 21, 2017