Distributed Quality of Service Routing

Abstract

The goal of QoS routing algorithms is to find a loopless path that satisfies constraints on QoS parameters such as bandwidth, delay etc. In distributed QoS algorithms, the path computation is shared among various routers in the network. These can be classified into two categories depending on whether the routers maintain a global state or not. Algorithms based on a global state information have less message overhead in terms of finding a path. However they have inherent drawbacks like routing with imprecise global state information, frequent message exchanges to maintain the global state etc. Hence such algorithms are not scalable. On the other hand, algorithms based on a local state information rely on flooding techniques to compute the path. Hence they have high overhead for finding a path. In this paper, we propose a distributed QoS routing algorithm, that maintains a partial global state and finds a path based on this limited information. Experimental results show that, the overhead of our algorithm is lesser than those that rely on flooding. The results also show that the impreciseness introduced does not affect the call admission ratio greatly.