Thursday, August 21, 2008

New Comcast Traffic Shaping Plan

Comcast has a new plan to deal with bandwidth hogs: slowing down their broadband access connections for periods of 10 minutes to 20 minutes, at peak congestion periods. Comcast hasn't yet offered a definition of what a "heavy user" is or how much bandwidth consumption qualifies one as a "heavy user."

At some level this is a marketing opportunity for fiber-to-home networks that should be able to operate without such restrictions. At another level, "heavy users" probably are not profitable customers at this point. Not only do their subscription fees not cover their consumption of network resources, but they also create problems for other users.

Still, the longer-term issue is that the usage profile of a "heavy" user today becomes more like the "normal" usage profile at some point in the future, when most people use the Internet to watch video.

Long term, there is but one reasonable alternative: much more bandwidth for every user. In the near term, there are some serioius marketing issues to grapple with. Though the overwhelming percentage of users will never encounter the traffic shaping, the existence of such shaping then becomes a potential marketing drag for some, an opportunity for others.

But it is a complicated matter. In truth, competitors probably are happy "heavy" users are Comcast's problem, as such customers are not profitable, and in fact create externalities. It isn't just the stress on the access networks. Comcast pays transit fees for all those video bits. So higher usage really does impose usage-based costs.

This will be an interesting marketing challenge for Comcast and its competitors. Longer term, it also is a packaging and pricing challenge, since most users ultimately will wind up consuming lots more bandwidth as video becomes a staple of Internet activity.

1 comment:

Anonymous said...

Great Traffic Shaping Plan!
This is a brilliant post thank you for publishing it!

Cheers,
Rondah

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...