NeTS-NR: Pricing, Shaping and Aggregation in Broadband Access

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

The motivation behind this proposal is to address a problem central to the fu-ture deployment of broadband, especially residential broadband, the problem of service pricing models and cost recovery. The focus of this research is on un-derstanding and guiding how the current pricing situation might change as the speed of residential broadband increases-if optimistic visions hold it might be 100 times faster in 10 years. First, current approaches to pricing, based on long term measurement and service tiers, may not be effective as the peak-to-average ratio of traffic changes. The second research issue is the "heavy user" problem-how to deal with the exceptional user that generates traffic loads an order of magnitude or more greater than the average users. The heavy user problem potentially worsens as the capability for high-end usage goes up by two orders of magnitude. The third issue is the "incomplete aggregation problem". In the core of the Internet today, much of the burstiness of traffic has been smoothed out, so providers negotiate contracts that approximate long-term av-erage traffic rates with some allowance for residual traffic bursts. As the peak to average ratio of individual flows increases, the residual burstiness in traffic ag-gregates of a given size may go up. A central question for this research is there-fore what is a rational model for the cost (and pricing) of carrying this bursty traffic. The goal of the research is to contribute to the development of the broad-band access, and to encourage the continued deployment of advanced broad-band services.