Template Tools

We've moved

Everything since June 1 is at  fastnet.news. This is just the archive from before June, 2015

Fiber News
Verizon to Buffalo: Drop Dead!
Written by Dave Burstein   
Tuesday, 03 May 2011 20:01
Sabine WomenVerizon FiOS is the best large network in the Western world, Ivan Seidenberg's enduring legacy. Ivan's decision to ccancel the last five million lines was a blow to the areas affected. In Buffalo, the second largest city in New York State, Mayor Byron Brown led a rally in front of the Verizon office demanding "Don't Bypass Buffalo: Verizon Build FiOS Now." Brian Meyer, Buffalo News via DSLR
      Rust belt Buffalo was the 15th largest city in the U.S. in 1950, but population has declined and moved to the suburbs. There's a commercial downtown and generally low rise but dense housing throughout the city. There are 145,574 housing units at an average density of 3,584.4 per square mile (Wikipedia,) which means a comparatively inexpensive fiber build. 
      Verizon continues to maintain they are not redlining poor and black areas, but the data are highly suggestive. They've brought FiOS to many of the affluent suburbs of Buffalo, but to little of the 40% black city. They also have excluded the 63% black city of Baltimore and the 40% black and Latino Boston, while upgrading nearby suburbs. In New York City, mostly white Staten Island is nearly completely FiOS while poor and black neighborhoods lag behind.
      Verizon is probably telling the truth that they did not choose to target FiOS at white areas. The actual record of Verizon - and AT&T - on issues of race is far better than most U.S. companies. But the criteria they did use, such as rate of business growth, had a clear effect. So did the decision in 2002 to stop investing in territories for sale. Verizon New England and the Verizon territories sold to Frontier had some of the lowest rates of DSL coverage in the developed world. That continues true in much of upper New York State, southern Virginia, and Pennsylvania, territories that Verizon is continuing to try to sell.
      Verizon treats these areas the way the early Romans treated the Sabine women. http://bit.ly/kDBkhg

Terabytes to Africa at $2-3/Broadband Connection
Written by Dave Burstein   
Wednesday, 20 April 2011 11:53
Africa_cables_Steve_Song_260500 gigabits of the West African Cable System landed at Yzerfontein in South Africa yesterday, with connections along the way to Namibia, Angola, the Democratic Republic of the Congo, the Republic of Congo, Cameroon, Nigeria, Togo, Ghana, Côte d’Ivoire, Cape Verde as well as the Canary Islands, Portugal and the United Kingdom. WACS is ultimately designed for 5 terabytes and higher.
     Back of the envelope, the $500M cost of the Alcatel-built cable equates to a landed price of $2-3/broadband customer for generous international capability. Although cable capacity is sold as 20 year IRU contracts, I think it's more realistic to spread the cost over 10 years. That's about $100K/gigabit/year, or just under $10K/month. Add $4K to purchase transit in London, and the direct operator cost is $12-16K/month/gigabit. Add markups and uncertainties, and the delivered cost/gigabyte to the large operator is $20-30K/month. Allocating 100 kilobits/customer - more than most networks average - spreads that 10K customers at $2-3/month.That's more than the $1/month typical in most of the U.S. or Europe, but no longer a great barrier to delivering a full Internet experience to most of Africa. The question now is whether the savings will be passed on to consumers.
     The problem: "Telco competition in South Africa is like pro-wrestling. There's a lot of shouting and posturing but the game is fixed." Song.
      A larger version of Steve Song's great map below and his presentation on the remarkable growth of fiber to Africa is worth a look http://bit.ly/hg9Wqf. The other map below, from the Ubuntu Alliance, shows the relative lack of fiber connecting inland Africa.
Prices Going Up As #1 Level 3 Buys #2 Global Crossing?
Written by Dave Burstein   
Monday, 11 April 2011 20:52

backbone_market_renesysBoth Roger Cheng at WSJ and Evelyn Rusli of the NY Times warn of higher backbone prices if the government allows the merger. There's no other logical reason to pay $3B for a company that has lost money for each of the last four years. Jim Crowe of Level 3 is claiming there will be $300M of "savings" even though the total SG&A was $431M. Much of that is sales and support that need to continue, I can't see anywhere for him to find a return like that except in higher pricing. (Global Crossing less of a backbone)

Pricing dynamics on the backbone are sufficiently complicated I put the question mark in the title. I haven't done enough analysis to join in the conclusion. It's almost definitely right to stop the deal but that's highly unlikely. One of the fortunate accidents of policy is that we've had enough players on the backbone to allow competition to do its magic. If that's no longer true, in the absence of strong regulation we'll see high prices and diminished Internet growth.

       Jim Crowe of Level 3 is one of the great salesmen of the era. He has a reality distortion field among investors although he's lost big money for a decade. One of Jim's problems is self-delusion. Long after Andrew Odlyzko had proven the Internet was not doubling every 120 days, Jim was touting that error to Wall Street. I think he believed it. 

Alcatel 10G GPON: Now on Sale
Written by Dave Burstein   
Sunday, 13 February 2011 16:02
Flexion_fiberVery few of the million or so installed GPON hubs have congestion problems so it's early to see much demand for 10 gig GPON. But when Huawei did a demonstration with Verizon of 10G, I believe Alcatel needed an answer. They now are the first to announce commercial availability of 10G GPON on their ETSI model. 

     GPON architecture has proven cheap and reliable, so carriers with GPON experience are using it for jobs like backhaul from remote DSL terminals and mobile cell sites. That's a niche market but an expanding one. For example, AT&T this year intends to run fiber to 15,000 cell sites and GPON is one of the architectures they've looked at. Extra capacity from 10G may be valuable there. 
5M Lines of Gigabit+ Fiber Home
Written by Dave Burstein   
Wednesday, 20 April 2011 18:48
Virgin_Media talktalk
Duncan Tait of Fujitsu vows to deliver 5M lines of gigabit+ fiber home to small town and rural Britain, perhaps the largest fiber build in the Western world this decade. It's about 20% of the country. Fujitsu will provide gear, build and finance the network. U.S. owned Virgin cable will be the anchor tenant, expanding beyond the urban half of the country it now serves. Super salesman Charlie Dunstone (Carphone/Talk Talk) will also sell the network, with Murdoch's Sky and Vodafone up for grabs.
     Nearly all this build will be in the half of Britain where British Telecom now has a monopoly. The economics of being the second network in an area are very different from being the third, overbuilding where there already is a telco and a cableco. While fiber overbuilders coming third to market have generally struggled, Virgin/TalkTalk/Fujitsu has much better prospects.
    Fujitsu, with support from Cisco, is fine with the relatively modest $200-$300/home subsidy offered. This team is betting $billions fiber is profitable over time based on the real world costs and experience of Verizon, NTT and others. Communication Minister Ed Vaizey is so enthusiastic he joined in the press release. Vaizey, to make the deal final, needs only to require BT to share existing ducts and poles at cost + a generous profit, standard unbundling terms.
Verizon: FiOS Speeds Do Little for Economy or Education
Written by Dave Burstein   
Monday, 18 April 2011 18:26

Broadband's effects on the economy are so small even careful studies can't prove them. There's even less evidence that higher speeds - 100 meg FiOS versus 3 meg DSL - have major economic impact. There are plenty of reasons to deliver a buffalogreat Internet for (nearly) everyone, but John Bonomo of Verizon is right to say  "it's unfair to suggest that the absence of FiOS in Buffalo is hindering its economy, education system and neighborhoods." 

   When even Verizon isn't claiming big economic benefits it's time for people like Drew Clark to stop making unproven assertions like "broadband is a key driver of job creation and economic activity." Drew is a friend and a good reporter who needs to look again at the evidence. Nearly every credible scholar agrees the payoffs have been exaggerated.

   Bill Lehr of MIT is the latest to speak out, particularly important because his earlier work is frequently cited to make job claims. "The results from my earlier research found there was a positive impact on jobs and other metrics of economic growth associated with expanded broadband deployment. But, that was based on the first generation of broadband deployment and we are not talking about that at this point.... its presence may not create jobs.

FiOS Beating Cable
Written by Dave Burstein   
Wednesday, 16 February 2011 11:32
shereff_bell_aliantCablevision added only 6K cable modem subscribers as FiOS is pulling ahead where they compete. If Verizon hadn't cut $2B from FiOS spending in 2010, Cablevision would probably have lost customers. "We have to get cable out of the home," Ivan Seidenberg told me years ago explaining FiOS. It didn't play out that way, as DOCSIS 3.0 and cable marketing strategies prevented Verizon pulling ahead. The slow rollout of FiOS also held them back.

      Bell Aliant in Eastern Canada accelerated their fiber rollout, adding 105K homes passed in 2010. The strong revenue increases in the fiber areas inspired them to plan 300K more homes passed in 2011, the largest North American fiber build after Verizon. Eastlink Cable had been hrting them badly and they decided they had no choice if they wanted to stay in the game.  Merrill Lynch's analysis confirmed this is the right move for them. Cincinnati Bell is nearly doubling fiber lines in 2011, to 159K. 
      Fiber is not dead.
40 Gigabit WDM PON from LG-Ericsson
Written by Dave Burstein   
Saturday, 12 February 2011 02:01
The original of this article  was a huge error based on my misreading the press release. This is a 1.25 Gig WDM PON, with 32 wavelengths, not  40G per wavelength. 10 GPON is nothing compared to the 320 gigabits (8 colors) or 640 gigabits (16 colors) promised by LG-Ericsson's 40 Gigabit "wave division multiplexed" PON, literally enough to serve a medium-sized country today.  Phil Winterbottom of Ericsson contends that kind of capacity "will become essential in applications like mobile backhaul for LTE," but each individual 40G wavelength is enough to support 200 cell sites at double the 100 megabits today provisioned.  Especially because 40 gig wavelengths typically require expensive "coherent detection" gear, the cost will be high enough this seems to me a narrow niche product.  I think this product is being announced "because they can" rather than because they expect to find a market. Sure is impressive, however. 

     WDM PON has been next year's technology for almost a decade, with the Korean government a long time supporter.  The technology is proven but they haven't been able to beat the remarkable drop in GPON costs. Outside of a government promoted 150K users at Korea Telecom, few systems have been sold.  Standard telco PON splits a single wavelength as many as 32 ways.  WDM PON dedicates a wavelength to each customer.  When Verizon began deployment of FiOS in 2003, they used BPON which was 512 meg down split 32 ways and even less on the upstream. Korea Telecom at the same time was considering a large WDM PON deployment that would have delivered a solid 100 meg in both directions. But Samsung and other vendors were not willing to bring the price down enough.  

<< Start < Prev 1 2 3 4 Next > End >>

Page 2 of 4