The End of the Flat Rate?

In 2008, a number of North American ISPs have publicly announced bandwidth caps, most prominently Comcast and AT&T. While this may seem as a step backwards for broadband customers at first sight, it actually is an improvement of the situation of the previous years with hidden and inconsistent caps. In fact, a true all-you-can-download plan across the entire customer base of an ISP at current price tags is commercially impracticable, despite the oft-lamented lack of a “real” flat rate by “consumer advocates”. Usually they try to drive their argument home by using everyday analogies such as monthly public transport travel passes, that also come with no usage limit. That analogy is of course totally flawed. Just as transportation authorities rely on the fact that physical presence in their vehicles is limited by natural means (Who on earth has allowed that camp site on our train?), every ISP has an oversubscription ratio without which no one would be able to pay for high-speed Internet access.

The question not really is why ISPs do this. For the sake of argument, they usually claim the so-called ”bandwidth hogs”, but even beyond services like P2P file sharing, the rising popularity and resolution of online video content shows which way we are headed on the bandwidth race track. The question should instead be, how to do it in a way that keeps their business profitable and customers happy and paying for their service. And exactly this — high-quality service — should be the focus of any bandwidth management solution. Bandwidth management should not be something that becomes allowed or accepted, but a service provided by the ISP to its customers.

So the question is: How can bandwidth management be implemented as a service to broadband subscribers and to the benefit of the ISP? There are many options.

Very simple prioritization schemes can be implemented to improve the performance of interactive applications such as Internet telephony, online gaming and Web browsing. This is even more important for congested networks commonly found at WiFi hotspots and hotels. How often did you try — and fail — to do a decent Skype call from your hotel room while on a business trip? The benefit for network users is clear. But how about the ISP?

Traffic Management

A major challenge for every ISP are diurnal usage patterns. Each network has to be provisioned to accommodate the daily data rate peaks, while the infrastructure remains idle during most of the day. With P2P making up over 50% of all network traffic and downloads usually lasting several hours, we have a good candidate for off-peak-hour traffic. And with a simple one-rule configuration in bandwidth management systems that assigns P2P a lower priority than everything else, we can make room for other, more important traffic during peak hours, but use all available resources if the network is not otherwise utilized. This approach effectively pushes P2P traffic to off-peak hours, which means a more even load distribution on network resources and — more importantly — a higher performance for non-P2P in times of high demand.

Many other, more elaborate schemes are conceivable that help to retain the flat rate models we have become used to. Whatever ISPs decide to do, they should clearly communicate their traffic management policy to their customers — and they should offer it as a value-add service.