A self-learning route selection scheme using multi-path searching packets in an OBS network

In an OBS network., some contention resolution schemes at a core router have been proposed. However, in terms of the optical device technology, it is difficult to implement the contention resolution schemes such as optical buffers at a core router. In this paper, we propose the scheme that reduces the probability of burst contention by controlling the route at an edge router without resolving burst contention at a core router. Each edge router learns a suitable route to the destination edge router autonomously by using newly employed feedback packets and search packets. Due to the self-learning at each edge router, the traffic load is distributed in an OBS network. Therefore, our proposed scheme can reduce the probability of burst contention. According to computer simulations, under nonuniform traffic, our proposed scheme can reduce approximately one decade smaller burst loss probability compared with the conventional shortest path routing method.