Concurrent bandwidth scheduling for big data transfer over a dedicated channel

Abstract

Many applications in various domains are producing colossal amounts of data, now frequently termed as 'big data', which must be transferred over long distances for remote operations. Such data transfer requires dedicated channels with high and stable bandwidth provisioned by high-performance networks HPNs through bandwidth reservation. For a data transfer task, from an application user's perspective, a common goal is to achieve the earliest completion time ECT, while from a network operator's perspective, a common goal is to achieve the shortest duration SD. In this paper, we investigate the problem of scheduling as many concurrent bandwidth reservation requests BRRs as possible over one dedicated channel in an HPN while achieving the average ECT and the average SD of scheduled BRRs. We show that both problems are NP-hard, and propose a heuristic algorithm for each. The performance superiority of the proposed algorithms is illustrated by extensive simulations in comparison with three other algorithms in terms of multiple performance metrics.