Resource allocation with advance reservation using artificial neural network in elastic optical networks.

The new technology of elastic optical network (EON) envisages a favorable solution for accommodating the heterogeneity of traffic over the future internet and to fulfill bandwidth-hungry requirement of today's world. Advance reservation (AR) is obligatory on initial-delay-tolerant services, like e-science and grid computing, that are having a remarkable involvement in next-generation internet traffic. In this paper, the authors proposed a routing and spectrum assignment (RSA) algorithm, which improves spectrum utilization by minimizing the blocking probability. This technique ensures spectrum availability for each connection request, but depending on the call arrival time and slot availability, there may be some delay and it is referred as spectrum allocation using reservation. The other proposed algorithm considers reservations based on prediction called as spectrum allocation using prediction and reservation. The simulation results show that the latter technique increases spectrum utilization even more and decreases call blocking compared to existing RSA algorithms. To the best of the author's knowledge, this is the first effort of spectrum assignment using prediction and reservation. [ABSTRACT FROM AUTHOR]