Scheduling for Time-Constrained Big-File Transfer Over Multiple Paths in Cloud Computing

Abstract

Enormous increase in the scale of big-file generated by cloud computing has brought many challenges to both computational and transfer infrastructures. For large file transfer in cloud computing, the scheduling with intelligent routing is indispensable to improve the efficiency of network resource, especially when file transfer over multiple paths is allowed. However, it is a non-trivial task to efficiently select a set of feasible paths for file transfer under a tight delay constraint. In this paper, we consider two kinds of file transfer scheduling problems: single-file transfer scheduling (SFTS) and multi-file transfer scheduling (MFTS), both of which require the file transfer to be conducted under a delay constraint. To address the SFTS problem, a heuristic algorithm together with an exact algorithm SFTS-A is proposed. Comparing with the exact algorithm, the heuristic algorithm takes less running time and is able to address the SFTS problem efficiently although the solution may not be optimal, which is adequately illustrated by simulation results. For the MFTS problem, we propose a heuristic algorithm MFTS-H with dynamic reconfiguration scheme for addressing the transfer scheduling problem for multiple concurrent files. The corresponding simulation results demonstrate that the proposed algorithm can efficiently support multiple file transfer over multiple paths in terms of delay and bandwidth utilization.