On the effects of allocation strategies for exascale computing systems with distributed storage and unified interconnects

摘要

The convergence between computing- and data-centricworkloads and platforms is imposing new challenges on how to best use the resources of modern computing systems. In this paper, we investigate alternatives for the storage subsystem of a novel exascale-capable system with special emphasis on how allocation strategieswould affect the overall performance.We consider several aspects of data-aware allocation such as the effect of spatial and temporal locality, the affinity of data to storage sources, and the network-level traffic prioritization for different types of flows. In our experimental set-up, temporal locality can have a substantial effect on application runtime (up to a 10% reduction), whereas spatial locality can be even more significant (up to one order of magnitude faster with perfect locality). The use of structured access patterns to the data and the allocation of bandwidth at the network level can also have a significant impact (up to 20% and 17% reduction of runtime, respectively). These results suggest that scheduling policies exposing data-locality information can be essential for the appropriate utilization of future large-scale systems. Finally, we found that the distributed storage system we are implementing can outperform traditional SAN architectures, even with amuch smaller (in terms of I/Oservers) back-end.