Monday, May 14, 2007

Is Enough, Enough When It Comes To Bandwidth AT&T?

Time to get back to telecom issues and leave content and services alone for a few articles.

Ah, the glory days when we proclaimed at AT&T that 56 Kbit/s was all the bandwidth a business would ever need.  A few years later we updated that cap to a T1 or 1.544 Mbit/s.  Lately, The New AT&T proclaims that 35 Mbit/s is enough bandwidth for your TV, phone, and Internet access.  Is it?

Comcast does not think so.  Earlier this week, Brian Roberts, CEO of Comcast, demonstrated Internet access speeds of 150 Mbit/s at The Cable Show in Las Vegas.  The demonstration utilized the DOCSIS 3.0 standard to bond together the equivalent of 4 analog TV channels to obtain the desired bandwidth for the demonstration.  They plan on reclaiming bandwidth on the plant to offer 100 Mbit/s Internet access to homes by moving to an all digital plant, utilizing switched digital video, reclaiming bandwidth from legacy applications, and implementing MPEG-4 compression according to Tony Werner, Comcast's CTO.  Comcast is not resting on its laurels in the triple-play arms race.

Additionally, Comcast will have to incorporate node splits to ensure enough bandwidth to each home.  Splitting a node costs between $3.35 to $26 per home passed depending on they way that they actually split a node.  About 65% of those splits are of the "logical" type that cost $3.35.  A small cost compared with some of the telcos upgrade plans.

Let's compare these figures to the cost of AT&T's Project Lightspeed.  Also this week, AT&T announced that the capital expenditure of Project Lightspeed would increase from $4.6B to $6.5B and the number of homes passed to 18 million; down a million homes.  That announcement increased price per home passed to $361 which is now about half the cost per home passed of Verizon's FiOS project that is delivering 50 Mbit/s Internet access alone.  To reduce capital expenditures, AT&T relies on a fiber to the curb approach and existing copper cable for the last thousand feet which limits bandwidth to the home.

In addition to the cost, AT&T depends on a completely switched digital video network to squeeze the most out of the limited bandwidth over their aging copper plant.  The Microsoft software that powers their IP-TV network has proven to be less than reliable to date.  Add this up:  a totally new switched digital video network, the limitations of a copper infrastructure, and a multi-billion dollar build-out for a network that is barely sufficient to meet today's triple-play needs. 

Contrast this risky strategy with Comcast's low cost, high bandwidth network and you would think that the market would be going crazy for Comcast stock.  The opposite is true.  AT&T's stock keeps rising while Comcast's is holding steady after a setback in January.  The market is fixated on short-term numbers instead of the long-term picture.

In the triple-play arms race, Comcast is better suited to capture a greater market share with faster Internet access, a greater selection of HDTV programming, more responsive VOD, cheaper voice services, and an integrated suite of packages ala Zimbra.  Despite of what Om Malik says, Comcast is delivering the innovation to make it a triple-play market leader.


  1. You are right on with your comments about AT&T facing challenges with respect to their choice of broadband technology.

    It is not widely known but their IPTV customers are limited to only 1 HD channel being viewed in the household at one time. With the significant reduction in HD TV prices, not too many people will be purchasing a SD set much longer. Imagine now that you get your new Vizio LCD TVs home only to discover that you can't watch the Spurs-Suns game in your den because your wife is watching Desperate Housewives in the living room!

    Verizon may be getting hammered for their investment in FTTH but in the long run they will have a much more competitive offering.

    Speaking of node splits, I was talking with an MSO excutive recently who said that it is their intention to get their homes passed to less than 50 per node. This combined with channel bonding, SDV and reclaiming analog channels with put the MSOs in a decent near-term competitive position.

    As time goes by the homes passed/mode will be reduced and the MSOs will eventually be forced into a FTTH model. AT&T will face this same investment but will lack the near-term flexibility of the MSOs.

    Should be fun to watch from the sidelines and as a consumer.

  2. I know of a couple of the large MSO that are looking seriously at PON systems. Seems that they have already anticipated this need.

    I would like to see munis support a third pipe to the customer with open access that would allow multiple providers to compete with the telco/cableco duopoly.

  3. There are two driver types for Bandwidth Growth; Organic which generally follows Moore's law; and IP Video requirements.

    The trend in data rates since 1980 has been speeds doubling every two years.

    1.5mb/s is on it way out,6mb/s will begin phase out 2012, 24mb/s will phase out 2015. Passed this rate you can no longer operate efficiently using anything but fiber. Metallic plant in the distribution even though longer in remaining life than in the feeder is limited to 7 years.

    Forecasts in the U.S. show consumer adoption of IP Video increasing from 20% of HH to 60% of HH in the next 5 years.

    GPON architecture most recently announced by AT&T will give them 38mb/s. In 5 years, this will put them at a competitive disadvantage.

    MSO's continue to work and manipulate their HFC plant inching fiber closer and closer to the homes as they reconfigure homes passed.

    Verizon got it right on this issue.

    They will be the first to market with availability ready for consumer adoption.