It didn’t take Comcast long to run away from its latest embarrassment. On Wednesday, Comcast Senior Vp Mitch Bowling told Bloomberg News that in an effort to control traffic, Comcast might slow down the transmission of packets from its heaviest users by “10 minutes to 20 minutes.” Here’s the story. PC Magazine had the same story.
By Thursday, Comcast had backed off, telling IDG that the company had made “no final decisions” on how to proceed with a traffic management plan.
The irony here, of course, is that the first stories on Comcast’s plans to “deprioritize” packets from its biggest users broke as part of the coverage of the Federal Communications Commission’s order smacking Comcast down for what the Commission called “invasive and outright discriminatory” practices that flew in the face of Federal policy to continue the development of the Internet. Comcast has 30 days to tell the Commission how it plans to put network management in place that doesn’t violate standard network protocols.
Apparently, at one point the plan was to delay the traffic of heavy users until, that is, someone at Comcast realized they had another PR disaster on their hands and backtracked faster than a Comcast user with supplemental Powerboost. You remember Powerboost – supposed to make it easier to download lots of material faster. It will be interesting to see how that enhancement survives in the new era of enforced Internet shortages.
More than Comcast’s obvious embarrassment, however, the announcement that it was even thinking about holding up the traffic of heavy users is just another demonstration of how far off the thinking is of those, like John McCain or FCC Commissioner Rob McDowell, who continue to insist there is a “free market” for broadband. In what other industry could a provider willy-nilly hold up the service of a customer using the bandwidth he or she is paying for without suffering any competitive consequences?
Not in the broadband market. The testing centers Comcast is using are, for the most part, small towns with little or no competition – Chambersburg, PA; Warrenton, VA; East Orange, FL (nor incorporated) and Lake City, FL – all have fewer than 20,000 people and in some cases a population with lots of low-income residents. The exception, Colorado Springs, CO, has 360,000 people and some competition that may eventually develop from two other cable systems, Falcon and Porchlight, which have been awarded city franchises to compete with Comcast. But at the moment they still serve small areas and Comcast is still the dominant player.
For an illustration on the malicious evil of duopoly, consider this story from upstate New York. The customer’s telephone company, Frontier (formerly Rochester Telephone), imposed a 5 GB/month usage cap – hardly enough for some movies and TV. He might switch, but the cable company is Time Warner, which is playing around with a cap of similar size.
A revealing quote from the story cites an internal Frontier email which says, “The growth of traffic means the company has to invest millions in its network and infrastructure, threatening its profitability.” That’s a shame for a company that on August 5 reported “Solid 2008 Second-Quarter Results with Strong High-Speed Net Additions.” In their financial news release, the company said data and Internet revenue continued to grow, and their capital expenditures were up 14 per cent from second quarter 2007.
Even so, it is much easier, after all, simply to clamp down on your customers than to expand the business. It’s easy to concoct an artificial network shortage in a less than competitive environment. That’s one interpretation. The other, from the story, is that it shouldn’t be hard for the company to keep up if it makes the right investments.
A less-than-competitive market can lead to all sorts of mischief. Usage caps, “deprioritization,” even a service which violates every tenet of Net Neutrality can all be “justified” in the name of managing a network in a time of shortages.
The public never gets to see what causes the shortages. There are no cases before the FCC or state regulators in which evidence has to be presented for public view. Every once in a while there might be a data request, as Canadian regulators did recently, finding little need for “throttling.”
No, the eventual answer should be something that recognizes the reality of the situation. Broadband now doesn’t present sufficient competitive choices for consumers, and the policies should be changed to make certain that those choices can be created.
If they actually go through
If they actually go through with this, and a “heavy user” needs to call 911 via VOIP and can’t because of Comcast’s poor network management choice, I predict a very large bill coming Comcast’s way that they really will not want to pay.
NB: I am a Comcast customer
if the policy is
if the policy is “protocol-agnostic”, the question becomes what “user-agnostic” means … Comcast implies that because it would treat users the same regardless of protocol, then that treatment would comply with standard network protocols presumably consistent with net neutrality
but “heavy use” differs sharply in whether it applies to gegabytes or gigabits per second, and focusing on “heavy use in real time” in this case means gegabyte use, even though during peak periods, large numbers of “small users in real time” can also contribute to whatever congestion may occur, since all peak congestion arises from coincident use of gigabits per second - not total gegabyte use over time
therefore a policy which penalizes the heavy use of gegabytes over time regardless of protocol - even real time during peak periods - is not “user agnostic” in the sense that costs are not born in proportion to costs caused across all users - only the “heavy users” are singled out and penalized, in effect rewarding the remaining users with continued, penalty-free uncongested service
the recent suggestion by Vint Cerf to provide dedicated minimum transmission rates rather than capping gegabyte use would qualify as “user agnostic” because it would price bandwidth in proportion to cost caused across all users … the dedicated feature would shift risks to the provider where it belongs and discourage oversold capacity and artificial shortages, and “heavy use” would become a matter of what bandwidth tier is chosen by the user, regardless of how much of a particular tier is used over time
in this context, the notion that a small percent of users are using a large percent of the bandwidth is moot - if they are using more bandwidth, it’s been prepaid and secured, not available to others, so at the margin, whether they consume more or less units of bandwidth is agnostic in terms of how it affects other “smaller” users because all users would cause or avoid the same cost on an equivalent unit basis