On-line traffic contract renegotiation for aggregated traffic

Consider an ATM service, which is charged by the contracted peak rate, but where it is possible to renegotiate the peak rate against paying a small renegotation charge. Such a service represents a more flexible economic resource sharing than the current practice of fixed capacities. We present an on-line algorithm for renegotiating the peak rate that trades off peak rate charges, renegotiation charges and long-term delay. We compare the performance of the algorithm in terms of charge and delay with the optimal strategy employing full knowledge of the future traffic, and with the constant bit rate strategy. Experiments in a simulation and in an experimental environment with aggregated TCP/IP traffic over an ATM link show that the on-line algorithm considerably reduces the gap between the optimal and the constant bit rate strategy.