Toward Flexible and Fast Routing Strategies for Dynamic Network Provisioning

Abstract

Reserving bandwidth as needed in high-performance networks makes the fast and reliable data transfer with guaranteed performance possible in large-scale collaborative e-science. Besides the notification of acceptance or rejection for a particular reservation request, users normally want to know the earliest possible finish time or the minimum total transfer duration for the data transfer. Several routing algorithms have been proposed to achieve such desired goals given the data size, the data available time, and the deadline to finish the data transfer. Instead of directly processing the bandwidth reservation request (BRR) from users, our approach analyses various parameters to strategically narrow down the solution search space for fast system response. Adapted from some previous works, two algorithms are proposed to compute the reservation options with the earliest completion time (ECT) and with the shortest duration (SD) for multiple BRRs accumulated during a certain period. Extensive simulation results demonstrate the superiority of the proposed algorithms in terms of reduced execution time and improved success ratio of BRRs in comparison with existing scheduling algorithms.