Paper Title: Ceph and NVMe-over-Fabric: Fewer Servers and Higher Performance

Paper Abstract: Ceph protects data by making 2-3 copies of the same data but that means 2-3x more storage servers and related costs. It also means higher write latencies as data hops between OSD nodes. Customers are now starting to deploy Ceph using SSDs for high performance workloads and for data lakes supporting real-time analytics. We describe a novel approach that eliminates the added server cost by creating Containerized, stateless OSDs and leveraging NVMe-over-fabrics to replicate data in server-less storage nodes. It also dramatically reduces write latency to enable Ceph to be used for databases and to speed up large file writes. As part of the solution, we also describe how OSD node failover is preserved via a novel mechanism using standby stateless OSD nodes.

Paper Author: Taufik Ma, CEO & Co-Founder, Attala Systems

Author Bio: Taufik Ma is CEO and co-founder of Attala Systems, a cloud infrastructure company. He specializes in cloud infrastructure software and physical and virtualized servers. A business executive with a proven track record for commercializing disruptive data center technologies, he was previously Sr VP/Chief Strategy and Technology Officer at Emulex, where he grew the converged networking business to a $100M annual run-rate. Before joining Emulex, he was VP Marketing and Business Development at startup Aarohi Communications and GM Enterprise Platforms and Services Division at Intel. At Intel, he co-managed at $400 million+, 700-person division that was the fourth largest supplier of server systems worldwide. Mr. Ma holds an MSEE in Communication Systems from the University of Southern California (USC).