CircleID

This note reflects standard private sector economic analysis, but the Internet serves many public interests as well.  Ken Arrow and other public good economists have dealt at length with the differences, especially when the good involved has natural monopoly characteristics, such as transportation systems.  The access layer of Internet service is very close to a public good for most of us and needs policies appropriate to that fact.  Vint Cerf's recent commentary to that end is on point, for which the die hard "market at all costs" folks have given him a hard time.

As an operator of PAIX Internet Exchange, the notion that that broadband access providers are going to restrict the traffic from content providers who don't pay them is not operationally possible today, nor are they ready to support a commercial solution today that is viable for content providers to pay access networks. They don't even have back office infrastructure to support that and would take more resources to manage that to generate revenue at wholesale level.  Most of the access networks (Cable MSO, DSL, Wireless, etc) buy some form of upstream traffic from Tier-1 networks.  Content networks that I know would rather stuff it down the same path of the access network upstream than be be forced to pay, and make them hurt.  Can you imagine source filtering of content network at Gbps level on access network's upstream transit link?  To manage that at the router level would be an operational nightmare and will yield lots of customer complaints. 

ISPs and their broadband subscribers will be better served to follow the money flow of traditional content distribution model that has worked for nearly hundred years.  Trying to get contents to pay for bandwidth is muddying the water again in an endless spiral of bandwidth price war.  If it's a video, access networks should get a piece of the advertising revenue for each video served at the retail level.  This should be negotiated at commercial level.  Nickle and diming at bandwidth level by network engineers is not going to yield high margins for access networks, nor should that be a job function for engineers to generate revenue.  This is how it is done today, and it is not good for the Internet infrastructure industry.