Message Scheduling for Data Redistribution through High Performance Networks

With the emergence of large scale distributed computing, new problems bound to data transfers are appearing. We present the problem of data redistribution between two clusters connected by a high performance network. This problem consists in finding the best way to transfer data from the first cluster to the second one in the shortest possible time. In order to avoid slowing down the network, and the transfer, it is necessary to schedule the messages. This problem (named as KPBS) is known to be NP-complete and we present here a 2-approximation algorithm we developed.