Fast Net News, DSL Prime, DOCSIS Report

WiMAX and LTE will deliver about the same performance when both technologies are mature in a few years, perhaps as soon as 2011-2. Both are using the same advanced engineering techniques (MIMO, OFDM, etc.) and pure IP architecture. The best estimate I have of practical average speeds comes from Dick Lynch of Verizon, 4-10 megabits, although when we get actual results from large deployments the average speed results could surprise. In particular, all the wireless data models assume that few homes will watch HD TV over the net or run any other app that requires massive bandwidth. My first guess is that most folks wanting HD will keep a landline and there are reasonable ways to handle the traffic, but this needs to be tracked. The 70 megabit/70 miles claims touted widely are not for real deployments and deliverable average speeds.

WiMAX is working well now; LTE will be deployed in 2010 but the carriers - including Verizon - don't think it's ready for volume use until 2012-2013. Home receivers on WiMAX are still much more expensive than cell phones (>$100 for most) and LTE until 2012-2013 will probably also carry a premium. Nearly all the telcos in the West have decided to wait for LTE because it's designed as an upgrade for their existing networks. Terry Norman of Analysys predicts "a difficult future of WiMAX in the developed markets of Europe and North America;" WiMAX folks believe their 3 year head start will tip things their way.

4-10 megabits is faster than the majority of DSL connections in the world today, so clearly some people will go all wireless. We're seeing small signs of that in some markets, especially in the Mideast and East Europe if the incumbent prices DSL too high.

• The most obvious is persuading other parts of government to free up some space they aren't using, which is quietly underway.
• The FCC Technical Advisory Committee years ago enthusiastically recommended allowing “cognitive radio” or “software-defined radio” that tests the spectrum before transmitting. That will be a battle, but I'm told Jules has the courage for this, amid nearly universal support from technologists.
• Expand “white spaces” to other frequencies as well. The FCC is creating a “TV Bands Database” to authorize access to vacant TV channels on a market-by-market basis. Mike Calabrese points out “There is no reason to limit the functionality of the TV Bands Database to the TV band frequencies. If a potentially useful frequency band is not being used at particular locations (e.g., in New York City but not in West Virginia,)” make it available to others. In those bands, use “GPS and the capability to periodically check an online database of available TV channel frequencies in that discrete geographic location. TV band white space devices (WSDs) will be required to query a national database to determine available channels at their current location before transmit capabilities are engaged.”
• Use femtocells, as AT&T plans, to effectively double their spectrum capacity.
• Take seriously and enforce at renewal Jonathan Adelstein’s “Use It or Lose It” rule. Few in D.C. realize that wireless licenses are not automatically renewed, clear from reading some. The rules can easily be changed if the FCC chooses.  They could require covering 95-99% of the territory at a minimum speed and quality. Make capacity part of the buildout requirements. Or the U.S. can follow the lead of Canada and India and require a percent of revenues from the licensee, using the money for true universal services. (Blasphemy, but the right thing to do.)

T's network famously choked on the iPhone, and Verizon has a reputation as more reliable. But Darren Murph at Engadget tested the performance of U.S. mobile data and found that AT&T is typically delivering twice the download speed of the other carriers. His work seems reasonably careful; perhaps T's upgrades are finally doing their job. “AT&T's download rates obliterate the other guys. Seriously, it's not even close.” AT&T came close to 2 megabits down “across a variety of urban, suburban and rural locales.” Upstream rates were all around half a meg, and latency also similar, about 150 milliseconds.

Until Verizon deploys LTE, T will be much faster and they already are moving speeds higher. That means VZ will be behind most places until 2011 or 2012, then should jump ahead. Murph is cruel about the business side, calling the current 5GB plans “capped” (true) and preposterous. He also points out the limits of 4 player competition “In a fashion that only a colluder could love, the big four here in America all have matching monthly rate plans with matching monthly caps (5GB). So much for choice, right?”

I've previously reported that AT&T has made the strategic decision to deploy 5-10M femtos for a cloud across the United States, potentially doubling their effective spectrum. They are still not satisfied the technology works well enough, however, with interference issues remaining hard to solve in dense areas. I had the chance at the CITI anniversary event to ask Ivan Seidenberg of Verizon about femtos, and his eyes lit up at the opportunity. He wouldn't say anything beyond "our trials are very interesting,"

"The reality is that it is only government permission to use spectrum (licenses) that is scarce. Spectrum capacity itself is abundant. Indeed, while actual spectrum measurement studies are difficult to find, those in the public domain have demonstrated that even in the so-called “beachfront” frequencies below 3 GHz, the vast majority of frequency bands are not being used in most locations and at most times.
The gross underutilization of the nation’s spectrum resource should be an urgent concern for national broadband policy. Spectrum is not only an immensely valuable and publicly-owned resource, but one that is infinitely renewable: every millisecond that a frequency band is not used for communication, that capacity is wasted forever. ...

Mark McHenry, a former manager of DARPA’s NeXt Generation spectrum program, found that even in Manhattan and in Washington D.C. near the White House, less than 20 percent of the frequency bands below 3 GHz were in use over the course of a business day. McHenry’s NSF study demonstrated in a mix of urban, suburban and exurban areas that the vast majority of the most valuable spectrum bands are vacant or unused for the majority of the time. The highest occupancy rate on the prime beachfront spectrum below 3 GHz was just 13 percent in New York City, while the average across locations studied was just 6 percent. Across the country, this underutilized spectrum represents an enormous untapped capacity for broadband; particularly in rural areas where average usage of “beachfront” spectrum is in the low single digits.”

NAF also provided suggestions for creating maps of spectrum that could be dynamically changed and checked in real time by devices wanting to transmit. Michael Calabrese and Sascha Meinrath are doing important work, joined at this session by Kevin Werbach and Michael Marcus. http://bit.ly/VBr5x