Distributed routing based on an estimated input traffic matrix

An adaptive optimal routing strategy that is based on an estimated input traffic (origin destination) matrix is proposed. Due to variations of network topology and arrival traffic into the network with time, adaptive routing strategies have better performance than non-adaptive routing strategies. One of the best routing algorithms is optimal routing algorithm. To use optimal routing, certain information must be known. First the network topology and link capacities must be known. Second the input traffic matrix must be given. The input traffic matrix may be estimated. The proposed routing strategy may be used in each of two forms, centralized and distributed. In the centralized routing algorithms, every node sends its link flow (traffic) information to routing control centre (RCC) node, and then the RCC estimates the input traffic matrix. The RCC uses optimal routing to compute routing tables and sends them to all network nodes. The main advantage of this routing strategy is that it may be implemented in distributed form. In the distributed implementation of this routing strategy, every node distributes its link flow information to all network nodes and estimates the input traffic matrix based on the link flow information obtained from other network nodes. Then every node uses the optimal routing algorithm to compute its own routing table based on the estimated traffic matrix. The represented routing strategy involves a computationally intensive iterative procedure, and therefore within large networks, this routing strategy becomes impractical. The input traffic matrix estimation method is first introduced, then a distributed routing strategy that employs the estimated input traffic matrix and other local information is examined. A further application of this strategy is in network performance monitoring.