Availability of high-performance IO devices has led to the development of various IO interfaces/APIs. For many distributed data processing and storage systems that consider integrating high-performance networks (e.g.,RDMA) and storage devices (e.g., flash), the performance implications of the available options are not clear upfront. To provide guidance to application developers as well as system designers, we discuss if (or what) combinations of modern networking and storage stacks can deliver the full performance of flash storage to a networked application? We report our findings with regard to access latencies, peak performance, IO efficiency, core scaling, and we perform a comparison between a block and an RDMA interface to a flash device.
Keywords: Stüdi, Groß
By: Animesh Trivedi, Bernard Metzler, Jonas Pfefferle, Patrick Stuedi, Nikolas Ioannou, Ioannis Koltsidas, Thomas R. Gross
Published in: RZ3906 in 2016
LIMITED DISTRIBUTION NOTICE:
This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.
Questions about this service can be mailed to reports@us.ibm.com .