Performance Evaluation of RED-ONION: A High-Speed Disk-to-Disk Transfer System
TL;DR
Red-ONION is presented, a high-speed disk-to-disk transfer system that connects research facilities, on campus and beyond, to a computing center, and the end-to-end optimizations across the transfer software, the operating system, and the storage that this requires are described.
Abstract
Modern experimental instruments produce data faster than general-purpose file transfer interfaces can move it, so delivery to the computing infrastructure has become a bottleneck in the research process. At many universities and research institutes, moreover, the instruments that generate research data and the high-performance computing systems that analyze it are separated both geographically and organizationally, because each demands its own expertise and installation environment. Connecting the two seamlessly is a pressing challenge for data-driven science. This article presents RED-ONION, a high-speed disk-to-disk transfer system that connects research facilities, on campus and beyond, to a computing center. The system combines data transfer nodes, a dedicated high-bandwidth network, an all-flash parallel file system, and multi-threaded transfer software that parallelizes network transmission and storage access. The design targets the wire rate both along the entire path, from the read on the sender storage to the write on the receiver storage, and for a single file between one pair of nodes rather than only in aggregate over many files or nodes. We describe the end-to-end optimizations across the transfer software, the operating system, and the storage that this requires. We evaluate a prototype deployed over a 100 Gbps transpacific path between Atlanta and Tokyo with a 150 ms round-trip time, on which a single 1 TB file transfer reached 90 Gbps, delivering a terabyte in approximately 95 s. Moving a dataset of this size therefore becomes a routine step, and the computing center serves an instrument as if the two were co-located.