Comcast, the biggest cable operator in the US, has started testing a broadband meter in Portland, Oregon and plans to offer the service nationwide soon. The meter, essentially an application in the customer service portal, tells subscribers how many bits in a given calendar month they have "used", both for downloads and uploads. Developed internally by Comcast, the meter covers all devices connected to the broadband line. Every three hours the information is refreshed. The research bureau NetForecast confirmed that the meter works with sufficient accuracy.
Comcast maintains a monthly cap of 250GB, enough for downloading 120 films or 40 films in HD. The company claims that average use is 2-4GB per month, and in any given month almost 99 percent subscribers do not meet the cap of 250GB.
The meter will certainly meet a need. One of the arguments against 'usage-based metering' (payment based on data use instead of a fixed monthly fee) is that a customer has no idea how much data he consumes. In addition to data use, the system can provide more insight into the actual speed of the service (see our commentary 'Regulators should require ISPs to provide broadband speed check'). Let's not forget though that this is a perverted market. What other industry puts such restrictions on heavy users? These customers are often called 'bandwidth hogs', hardly a nice description for your best customers. Comcast previously resorted to cutting off heavy users, after the FCC forbid traffic intervention at the protocol level. As a result, Comcast could no longer direct peer-to-peer traffic (the white elephant in the room and one of the causes of the 'problem'), and the company had to come up with something at the customer level.
The big question remains does Comcast (and other ISPs) really have a problem with capacity, and if so, where: in the access network, the backbone or somewhere in between? Telecompaper's latest research report 'Dutch Broadband Speeds' (Q3 2009) shows that ISPs largely do not meet their promises: average download speeds are well below the 'headline' speeds advertised. That also applies to 'dedicated' FTTH networks, where broadband subscribers do not share bandwidth with the neighbours. In other words, if there's a capacity problem, then it's not in the last mile, but somewhere higher in the network.
This brings us to the underlying question: why don't providers do something about the backbone problems? Of course it's first a question of costs, as they are continually confronted with high traffic growth. Second, the problem can never be completely resolved, as there are numerous other bottlenecks, such as network equipment and household cabling. But the conclusion is clear: this is not something the market will solve itself. This is not per se a disaster – unless the government find that consumers and economic growth are suffering as a result.
If we go a question deeper, why doesn't the market solve the bandwidth problems? The most probable answer is that the market values the scarcity created. Scarcity means a certain price level can be maintained, and ISPs can avoiud being reduced to the infamous 'dumb pipe'. The market also appears unable to solve this problem, until some newcomer comes at the right moment with a business model based not on price but on volume.
Conclusion: Comcast appears to be offering the consumer a service, but is actually diverting attention away from the real problem: insufficient capacity in he backbone.
