Scheduling Algorithms on Optimizing Spatial and Time Distribution of Bandwidth Utilization

摘要

A multi-service network that can accommodate various services, such as enterprise VPN services and mobile services, is a promising solution for reducing OPEX and CAPEX. In this type of network, bandwidth can be used more efficiently by managing paths with respect to each time slot, because the time distribution of the bandwidth required by each service varies. In this paper, path selection algorithms for two types of requests are proposed. One is an algorithm for request with specified bandwidth (RSB) that calculates paths considering the bandwidth variation for different time slots, and the other is an algorithm for request with specified data volume (RSDV) that calculates paths considering the data volume. Experimental results showed that with the proposed algorithm for RSB, the network was able to accept 1.7 times as many requests as the minimum hop algorithm and 1.1 times as many as the exponential algorithm, which does not consider the time distribution of the bandwidth utilization. With the proposed algorithm for RSDV, the network was able to accept 1.6 times as many as the minimum hop and 1.5 times as many as the exponential algorithm with randomly selecting the transfer start and end time slots.