Multiple bulk data transfers scheduling among datacenters

Bulk data migration between datacenters is often a critical step in deploying new services, improving reliability under failures, or implementing various cost reduction strategies for cloud companies. These bulk amounts of transferring data consume massive bandwidth, and further cause severe network congestion. Leveraging the temporal and spacial characteristics of inter-datacenter bulk data traffic, in this paper, we investigate the Multiple Bulk Data Transfers Scheduling (MBDTS) problem to reduce the network congestion. Temporally, we apply the store-and-forward transfer mode to reduce the peak traffic load on the link. Spatially, we propose to lexicographically minimize the congestion of all links among datacenters. To solve the MBDTS problem, we first model it as an optimization problem, and then propose a novel Elastic Time-Expanded Network technique to represent the time-varying network status as a static one with a reasonable expansion cost. Using this transformation, we reformulate the problem as a Linear Programming (LP) model, and obtain the optimal solution through iteratively solving the LP model. We have conducted extensive simulations on a real network topology. The results prove that our algorithm can significantly reduce the network congestion as well as balance the entire network traffic with practical computational costs.