Enabling active storage on parallel I/O software stacks

Abstract

Abstract-As data sizes continue to increase, the concept of active storage is well fitted for many data analysis kernels. Nevertheless, while this concept has been investigated and deployed in a number of forms, enabling it from the parallel I/O software stack has been largely unexplored. In this paper, we propose and evaluate an active storage system that allows data analysis, mining, and statistical operations to be executed from within a parallel I/O interface. In our proposed scheme, common analysis kernels are embedded in parallel file systems. We expose the semantics of these kernels to parallel file systems through an enhanced runtime interface so that execution of embedded kernels is possible on the server. In order to allow complete server-side operations without file format or layout manipulation, our scheme adjusts the file I/O buffer to the computational unit boundary on the fly. Our scheme also uses server-side collective communication primitives for reduction and aggregation using interserver communication. We have implemented a prototype of our active storage system and demonstrate its benefits using four data analysis benchmarks. Our experimental results show that our proposed system improves the overall performance of all four benchmarks by 50.9% on average and that the compute-intensive portion of the k-means clustering kernel can be improved by 58.4% through GPU offloading when executed with a larger computational load. We also show that our scheme consistently outperforms the traditional storage model with a wide variety of input dataset sizes, number of nodes, and computational loads.

A Checkpoint of Research on Parallel I/O for High-Performance Computing

“…The active storage approach is discussed by Piernas-Canovas and Nieplocha [83]. Son et al [92] discuss how active storage can be implemented in parallel file systems without information about data layout and striping.…”

Active Flash: Out-of-core data analytics on flash storage

“…were demonstrated in this environment, the computational power of disk-resident controllers has not increased significantly since then, while the real-time demands of mechanical control make for a poor environment for user programming. More recently, active storage concepts have also been pursued in the context of parallel file systems [38], [46], harnessing the computing power of the storage nodes, or hosts dedicated to disk management and data transfer (e.g. Jaguar's Lustre parallel file system at ORNL uses 192 dual-socket, quad-core, 16 GB RAM I/O servers [2]).…”

DataMods: Programmable File System Services

“…Recent work [12], [13], [14] integrates active storage with object-based storage systems, and protocols such as iSCSI and T10, but are low-level facilities for arbitrary code injection. Son et al raise the abstraction level of active storage up to that of MPI-IO [15], allowing new MPI calls to trigger colocated computation.…”