Hierarchical source routing using implied costs

Based on a loss network model, we present an adaptive source routing scheme for a large, hierarchicallyorganized network. To represent the “available” capacity of a peer group (subnetwork), we compute the average implied cost to go through or into the peer group. Such implied costs reflect the congestion in the peer group as well as the interdependencies among traffic streams in the network. We prove that both a synchronous and asynchronous distributed computation of the implied costs will converge to a unique solution under a light load condition. Furthermore, we present a more aggressive averaging mechanism that, with sufficient damping, will converge to a unique solution under any traffic conditions. One of the key features of this paper is an attempt to quantify routing “errors” due to inaccuracies caused by aggregation. In fact, our experimental results show that these approximations are reasonably accurate and our scheme is able to appropriately route high level flows while significantly reducing complexity. In addition, we show how on-line measurements and multiservice extensions can be incorporated into the routing algorithm.