Fast Net News, DSL Prime, DOCSIS Report

Cisco and Washington estimates much too high. The almost unbelievable 100% per year growth in wireless traffic is over, with the growth in 2012 down to 69% in 2012 compared to 123% in 2011. The data is from CTIA, the wireless association, and is based on direct reporting carriers representing over 90% of U.S. subscribers. It's definitive, and corresponds to the trend I reported last year. 

   The growth rate drop will continue as the once in a generation surge due to the introduction of smartphones is winding down. I'd expect 2013 to fall under 50% but that's back of the envelope thinking, not analysis.

    40-50% growth is still pretty substantial, but the technology is keeping up. LTE Advanced, starting to deploy, provides 5-10 times the capacity of today's LTE. There's now no doubt the spectrum crisis was wildly exaggerated.

   I reported this last year as 2012 U.S. predicted 259 PB/month, actual 222 PB/month. World predicted 1,252, actual 884. The 100% growth rates for mobile data are disappearing, as predicted since 2009 by most experts and ignored by JG and other policy people. 86% in 2012 and a predicted 70% in 2013 remains high growth. There's an even lower number using some revised Cisco data - 62% instead of 86% - that I need to check out. The latter is closer to the CTIA data reported by Tim Farrar. Of course, even the 42% Cisco predicts for 2017 will require first rate engineering to serve, but the “crisis” was an invention of D.C.

   There is a hump in demand as people get smartphones and 3G/4G. That began early in the U.S. with the iPhone but is running its natural course. More than half the U.S. has smartphones so that growth is tailing off. Craig Moffett notes the same phenomena on the financial side. Smartphones have lifted revenue but the effect is playing out.

    As carriers start metering data, people learn not to watch video on their phones outside of WiFi range. Glen Campbell of Merrill Lynch predicted back in 2009 that pricing would cut demand and that’s kicked in with a vengeance. Cisco sees 46% of smartphone traffic offloading to WiFi by 2017. (Also predicted by Campbell in 2009.)

2012 U.S. predicted 259 PB/month, actual 222 PB/month. World predicted 1,252, actual 884. The 100% growth rates for mobile data are disappearing, as predicted since 2009 by most experts and ignored by JG and other policy people. 86% in 2012 and a predicted 70% in 2013 remains high growth. There's an even lower number using some revised Cisco data - 62% instead of 86% - that I need to check out. The latter is closer to the CTIA data reported by Tim Farrar. Of course, even the 42% Cisco predicts for 2017 will require first rate engineering to serve, but the “crisis” was an invention of D.C.

   There is a hump in demand as people get smartphones and 3G/4G. That began early in the U.S. with the iPhone but is running its natural course. More than half the U.S. has smartphones so that growth is tailing off. Craig Moffett notes the same phenomena on the financial side. Smartphones have lifted revenue but the effect is playing out.

    As carriers start metering data, people learn not to watch video on their phones outside of WiFi range. Glen Campbell of Merrill Lynch predicted back in 2009 that pricing would cut demand and that’s kicked in with a vengeance. Cisco sees 46% of smartphone traffic offloading to WiFi by 2017. (Also predicted by Campbell in 2009.)

Verizon and AT&T dominate postpaid, but prepaid is taking a share. VZ & T effectively manage a cartel, threatened only at the edges by prepaid carriers. Sprint and T-Mobile are falling further behind by about 1m subs per quarter, but rarely break cartel pricing. Verizon punished Sprint the last time they tried.
   Despite continually falling costs as wireless technologies improve, the advertised package price remained about $40 from 2007 until 2012. Prices had been coming down for a decade, bringing the U.S. to some of the best pricing in the developed world. Then Newtel and the old AT&T Wireless were swallowed up, reducing competition at a cost to consumers of billions each year. Extra fees - especially ever increasing text charges - mount up. Verizon’s 2012 pricing, followed closely by AT&T, raises the prices to an effective minimum of over $80, with minimum data. Verizon’s CFO Fran Shammo has told the street he intends to continue raising prices. Almost all choices lock you in for two years, although not long ago one year contracts were common.
      Prepaid alternatives are 30-75% cheaper in a much more competitive market. iPhones are finally available - at $650 - and the LG and Huawei smartphones for $100-200 are getting better. The price difference is so large that postpaid growth turned negative in Q1 as smart consumers move to prepaid. While some of the prepaid are smaller carriers (Leap/Cricket, MetroPCS) most are MVNO resellers of the big four. Sprint is using their Virgin and Boost Mobile prepaid brands in attack mode, and T-Mobile after the sale fell through is developing similar.
      Q2 2012 saw zero growth in postpaid smartphone sales over the prior year, but 91 percent growth in prepaid smartphones.  Verizon and AT&T have numerous strategies to hold back competitive prepaid growth deployed both in their offerings and allies like Tracfone. I carry a $12/month Tracfone (Telmex), a nice feature phone with about 100 minutes. it runs on the verizon network and has been just as reliable as the $50 service. People are starting to realize that if you don’t need a smartphone, it’s totally dumb to go the expensive route. The smartphones are coming along.
   The great fear in policy circles - including the FCC Chairman - is that VZ and AT&T pull away so far they can charge whatever they want and set almost any rules they like. That’s been at least partially true, evidenced by the high prices. The U.S. in 2004 had some of the best prices in the developed world. Since Nextel, AT&T Wireless, and Alltel were bought out, the competition weakened and U.S. prices are now higher than many.
   It’s too early to tell whether they have gone too far with price increases.

Often 3 meg when no other technology works. Ken Garnett at Cal.net has been serving heavily forested Northern California since 2006, forested country where many couldn’t get DSL. His Wimax/WiFi network couldn’t reach many prospective customers because of hills and trees. His new “White Space” network, using lower frequencies, generally solves that problem.

  As Dave Farber taught me a decade ago, there’s massive unused spectrum in most of the United States. Even in “crowded” space like New York, the majority of spectrum is unused. In Ken’s rural California territory, he tells me 90 MHz - 15 television channels - is typically unused. Garnett adds “We see that continually with newly transplanted residents in our service area who call us desperate for something, stating something along the lines of 'when I bought my house and moved here from [wherever], it never occurred to me that [cable/DSL] was unavailable.'" 

Unused TV channel space is the first target of “White Spaces,” which naturally can extend to other frequencies as well. “White spaces throughout the spectrum” is a key part of the technologist’s plan for wireless. Jim Carlson, a pioneer in this technology, is providing the equipment. The base station serves from 470 to 796 MHz in the international version, less to keep out of the way of 700 MHz LTE in the U.S. version. That requires a customized baseband modem, two million gates of Xilinx FPGA, as well as other components to be “frequency agile.” The CPE, currently fixed only, incorporates an optimized antenna. Google is using Carlson Wireless gear in their African trials as well.

  Current U.S. versions of Carlson’s RuralConnect peak at about 12 megabits, with 3-6 megabits the typical customer speed. Jim tells me that will double in a few months when he supports “bonding” two channels. Garnett confirms that in extreme rural areas like his, two adjacent channels are generally available. In a year or so, Carlson plans to bond four channels for at least another doubling of speed.

  Costs remain high because this is currently customized, low volume design. CPE costs about $600/home, base stations about $200/home if they serve 15-20 customers. Carlson’s confident he’ll keep lowering the price as his volume builds. I’m told white space equipment is not terribly more complex than WiFi and in large volume would have similar cost to WiFi. Chipmakers haven’t yet jumped in, however.

  The “spectrum shortage” is mostly a myth, with both Craig Mundie of Microsoft and Eric Schmidt of Google affirming the issues can be solved with different policies.

    Here's the pr

Big names don't make top 15. The Wireless Intelligence rankings makes clear the market has shifted to Asia and a few large groups. http://bit.ly/TrNxEC . By sales, however, Verizon and AT&T are #2 & #3. Only some of that is the high U.S. percentage of smartphones and take rate on data. U.S. prices are yet again marked as the highest. At http://www.nytimes.com/2012/10/15/technology/americans-paying-more-for-lte-service.html, Kevin O'Brien reports GSM Association data that Americans are paying twice as much as Europeans. Important data below.