IORE: A Flexible and Distributed I/O Performance Evaluation Tool for Hyperscale Storage Systems

Abstract

The process of input/output (I/O) performance evaluation, when considering the emerging data-intensive applications, over modern multilayered parallel I/O software stacks and hyperscale storage systems, has proved to be a complex effort. Although the importance of this subject and existence of common tools, the I/O research community still searching for a unified, user-friendly, and flexible approach to tackle the well-known issues. Proposals such as application I/O kernels and synthetic I/O workload generators have been widely adopted for the I/O performance evaluation. Toward to meet a differentiated mechanism, we propose the IOR-EXTENDED (IORE) research work, which can be considered as an enhanced I/O performance evaluation tool. More than a synthetic workload generator, IORE was designed to reduce the time to response in I/O performance experiments and analysis. In this paper, we present the IORE environment, which indicates from the first results a promising experimental tool.