Greg Schulz

Register to Attend this BoF

Many server storage and I/O problems are often the result of host CPU and associated software bottlenecks. This panel explores the challenges, issues, and benefits of addressing NVMe over TCP deployments without compromise. The session will explore server, storage and I/O workload testing techniques, tools, methodology and approaches to show NVMe over Fabrics including TCP can be accelerated, while freeing up host CPU resources for other software defined workloads. With a continued shift towards software defined storage, networks and higher level applications, host CPU cycles are in more demand. Boosting application performance, efficiency, and effectiveness of server CPUs including reducing overhead are key priorities for legacy and software defined datacenter environments.

Greg Schulz

Independent Industry Analyst & Author

StorageIO™

Bob Dugan

Director of Engineering

Chelsio Communications

NVMe over TCP has emerged to provide a more powerful, and compute resource friendly NVMe-oF technology deployment option. Not only does NVMe over TCP help with lower deployment costs and reducing design complexity, it also enables higher performance with lower latency. In essence, NVMe over TCP extends NVMe across the entire data center (from on-prem core to edge to cloud) using the simple, standard, and efficient well known TCP/IP fabric networks. This panel explores the challenges, issues, and benefits of addressing NVMe over TCP deployments without compromise. The session will explore storage and I/O workload testing techniques, tools, methodology, metrics and approaches to show how NVMe over TCP including TCP/IP and NVMe processing can be accelerated. Likewise, this panel discussion will look at how NVMe over TCP frees up host initiator as well as target resources to improve application performance, and server storage I/O workloads including compression, deduplication, encryption, rebuilds, and tiering among others.

Greg Schulz

Independent Industry Analyst & Author

StorageIO™