Sunday, April 4, 2010

Broadband and the Wrong Numbers

The Economist has published an article extolling the Google broadband effort as well and commenting on the state of broadband in the US today.

There are three issues upon which I will opine.

First, there is an explosion of real time video applications which are being delivered across the Internet.

From the high end Cisco Telepresence to the lower end Skype, iVisit, and others. At the top end one needs to have a special room, high cost monitor and if a commercial entity you need to have access to a network specifically made for Telepresence, namely a high cost ATT network. This is a clear barrier to entry for major players. But alas that is ATT and Cisco. Universities are disintermediating this by using the Internet and specifically Internet II.

This video explosion, real time video, is changing the way we use and see the Internet. I use it as a real time displacement for my office, for conferences and the like. It is low costs and works well. The problem is that at the low end it is designed for the least common denominator, namely DSL. The result is poor images and the like. But the potential is there to explode.

Second the Economist misrepresents the true costs of deploying fiber. This is a pandemic problem amongst policy planners which is why frankly centralized Government planning is a fundamental evil in and of itself. It is done by people who have no real experience.

The Economist wrongly states:

Though it cost Verizon a hefty $1,350 to bring fibre into a home, the $19 billion investment in FiOS was considered worthwhile. It stemmed the tide of telephone subscribers defecting to cable companies, which offered faster internet access bundled in their television and telephone packages at the time.

This is false. As we have demonstrated before many times the cost is closer to $3500. Just the ONT, CPE, and drop cost $1500! Then you have to add the fiber plant plus all the other stuff. Where the Economist got this number is any one's guess but it is not even close!

Fiber is costly, no matter where you deploy it. Yet it has certain long term strategic advantages that exceed the limited ability of cable. The Cable companies extol the virtues of DOCSIS which is trying to put ten pounds in a one pound bag, no matter what Comcast law suits try to portray.

Third, wireless is good but!

The Economist states:

In future, presumably, Verizon will offer customers who missed out on FiOS a chance to get high-speed broadband over the airwaves instead of through an optical pipe. The LTE (long-term evolution) technology that Verizon and other wireless carriers are testing is vastly cheaper to install than fibre. And with tweaks, LTE should be able to offer data rates of 150 megabits a second or more—much the same as the DOCSIS 3.0 broadband technology being deployed by the cable companies.

Let's deal again with some facts, something I find many time eludes the Economist. Well it is the Brits any way, they always seem to play fast and loose with facts, just look a Global warming, but a tale also for another day.

The facts are that LTE allows more efficient use of spectrum. Namely by using high modulation efficiency along with integrated multiple access via OFDM one can get close to Shannon limits on the bps/Hz. We show this below:

First, we show bps/Hz versus range for a variety of frequency bands. Remember that if we say have 20 MHz of bandwidth and we use a bandwidth efficiency of 7.5 bps/Hz then we have a maximum capacity of 150 Mbps! That is somewhat unrealistic but close. It makes the point. The lower the frequency the greater the range. Missing from this simple analysis is the complex issues of shadowing, diffraction, multipath and the like. For anyone interest you should read my papers in the early 1970s at MIT on my MIT site, I did most of the original research in the experimental area, thus I speak from experience, again not somethings which seems to bother the Economist.

Second we show the same plot but now as a function of frequency at various ranges.

Third and finally we show the bandwidth efficiency versus range for the 1.9 GHz band.

Note that we can still get 1 bps/Hz at 5 miles, assuming all we assumed above, but then it goes to zero. This means lots of cells.

Thus can wireless surpass cable, possibly, at a lower cost, maybe, but alas like cable the wireless systems must share the spectrum with others, something that is not necessarily part of fiber.

As for the FCC Report, here the Economist seem to be spot on. They state:

The report—all 376 pages of it—has, however, been greeted by a resounding raspberry. It calls for 100m households to get 100-megabit access by 2020, while the rest of the country is assured at least four megabits a second. Americans, in other words, will have to wait another decade to get what other wealthy countries now take for granted.

And perhaps this is why Verizon is chilling on its fiber deployment. The FCC is positioning itself at the market maker for broadband, another attempt to control the economy in a Galbraithian fashion by the current Administration.

As CNET states:

Verizon Communications is nearly finished building its Fios fiber-to-the-home network. And now it will concentrate on expanding its customer base, say executives.

Verizon, which began building its all-fiber network nearly six years ago, is quickly approaching its goal of passing 18 million homes in about 70 percent of its original customer footprint by the end of this year.

Verizon took a bold risk when it decided to build the Fios fiber network. Its telecommunications counterpart SBC Communications, which is now AT&T, decided to invest substantially less capital to extend fiber to the node or to the neighborhood. AT&T, which uses existing copper lines to deliver service to customers, said it would invest $4 billion in upgrading its network. Verizon committed to spending $22.9 billion.

Initially, Wall Street was skeptical about Verizon's bet. It costs Verizon about $750 per customer to wire an entire neighborhood for the Fios Fiber service. And for every customer who signs up for service, Verizon spends an additional $600 to bring wire directly to the home.

Why should Verizon spend billions when on the one hand the FCC is going to tell it what to do, on the other hand the Government is taking money way which is could have invested in infrastructure as a new health care tax. The uncertainty in the slowly crumbling free market is perhaps the canary in the broadband tunnel.

The Wall Street Journal stated:

Verizon Communications Inc. will no longer seek new cities to roll out its FiOS TV service as the company nears the end of its $23 billion network upgrade project.

The New York-based telecommunications provider has wagered heavily that its bundle of faster Internet service and television would give it an edge over its cable rivals, and has spent the last six years replacing much of its older, slower copper lines with faster fiber-optic ones. One of the hurdles has been getting community approval to offer television service, or what's known as a video franchise.

Verizon made it known that it would not seek anymore franchises after sending a letter last week to city officials in Alexandria, Va., telling them that the company has stopped seeking nationwide permission to offer television service in new markets. Among the other large cities not getting access to FiOS TV are Boston and Baltimore.

The company, however, said it is still working through existing negotiations with about a dozen local communities, mostly in New York, Massachusetts, and Pennsylvania. The company also continues to expand the availability of FiOS to its existing markets.

The true moral of the Fios tale is stated in this story. It is the Franchise. We wrote on this five years ago after we saw the irrational, in my opinion, stand by certain towns in New England where they sought the world from any new entrant when all we asked for was equality. But that was we believed then and now a cable induced action.

The Verizon target is:

Verizon still intends to connect 18 million households to FiOS by the end of the year, with some residual deployments slipping into 2011. By the end of the rollout, 70% of Verizon's households will have access to FiOS Internet, with slightly less getting access to TV service, Kula said. At the end of last year, 15.4 million households were connected to the high-speed network, although not everyone was able to get the service yet.

The 15 million HH connected at a cost to date of $23 billion is incorrect as we calculate it. One should recall as we have above the cost of the ONT and two CPE units alone is $1500. This the $1500 that the WSJ quotes accounts only for the total customer premises equipment and somewhere else there must be hidden the fiber and other electronics!

The Franchise is and will remain the cable companies most powerful ally. In addition is the cable boards or state commissions which are packed with political hacks in my opinion and experience and tend to keep the status quo. The last statement in the WSJ article is the most telling. Especially since we told them that six years ago! Namelky:

Achieving franchises on a community-by-community basis is among the arduous tasks. Getting approval often means meeting individual demands from each city, including setting aside channels for local use and other city improvements. It takes roughly 18 to 24 months to get a franchise, Mr. Kula said.

Thus why is the US "behind". Not due to any national policy, due solely to the Franchise, which exists no where else in the world. I had dealt with Franchises when I was at Warner thirty years ago. It was chaos then. Now it is reduced to stupidity. It it not even recognized by the FCC! They seem at the FCC to be totally clueless. It reminds me of 1968 and a "friend" who a summer vacation in Thailand and having them recount the tales of Vietnamese traffic on the Trail and then having them name the NVA rest stops after those on the New Jersey Turnpike. The JCS Staff apparently did not find such a naming as funny and the person who did it, albeit done with some sarcasm I gathered at the time, yet to those in the field at the time, the truth and the facts did resonate with reality. Thus the FCC in its "Plan" is akin to the JCS planning staff in 1968, they seem to get the "facts" by listening to themselves and failing to gain the truth from the trenches.

Does Verizon have a chance forward, say in wireless. There is still the Vodaphone problem. Also there is what I would term the Seidenberg problem as well. Not that Seidenberg has any problems it is just that he will be approaching retirement and one wonders what his replacement will be like, for the very survival of the company may depend on this person. Seidenberg brought Verizon to its current position, but like Moses may not be there when it enters the promised land, one devoid of copper and filled with fiber and wireless!