Get the Fastest Access to NVMe Storage

I am so pleased to see the momentum for Intel® Rack Scale Design (Intel® RSD) keeps building. In a recent blog I shared Intel’s work that enables dynamic composition and provisioning of Intel RSD compatible hardware in the OpenStack environment. Now I want to highlight Intel RSD storage solutions that use extremely fast, highly scalable NVMe over Fabric* interconnect technology. The best part is that you can see it for yourself at the Gartner IT Infrastructure, Operations Management and Data Center Conference* in Las Vegas (Dec. 4-7).

NVMe Overwhelms Shared Storage Solutions

SSDs based on Intel® Optane™ and Intel® 3D NAND technologies provide new levels of performance and price/performance versus traditional hard disks and NAND SSDs.  It’s now possible to employ NVMe* in multi-tier, hot-warm-cold storage strategies to economically and dramatically increase the performance of applications like big data, machine learning, genomics and multi-physics analysis and simulation, which require fast access to huge amounts of data.

The potential weak link in this scenario is the interconnect system used to access shared storage. Methods like iSCSI and SAN are giving way to faster technologies including NVMe over PCIe and NVMe over Fabrics. Using NVMe protocols with switched PCIe is the cheapest and fastest way to enable composable systems based on the Intel RSD architecture, and may be optimal for many scenarios.

For data centers that want shared storage that can span many racks, using NVMe over Fabrics via the Intel RSD architecture is a high-performance alternative. The NVMe over Fabrics protocol works over Ethernet and employs RDMA (remote direct memory access) to minimize latency and maximize bandwidth. NVMe over PCIe is supported by Intel RSD today, and NVMe over Fabrics support will be released in mid-2018.

The Intel® RSD Architecture supports both NVMe over PCIe (now) and NVMe over Fabrics (mid-2018) solutions with a full set of RESTful APIs.

See NVMe over Fabrics Demo at Gartner Data Center Conference

To demonstrate how Intel RSD works with NVMeF to enable a disaggregated, composable infrastructure, Intel and Supermicro* will demonstrate how to compose logical nodes on Supermicro RSD* products. The demo will run on remote Supermicro RSD BigTwin* 2U 2/4-Node Servers and Supermicro RSD Pooled NVMe storage units with 32 SSDs.Supermicro RSD Pooled NVMe storage units with 32 SSDs

remote Supermicro RSD BigTwin* 2U 2/4-Node Servers

At the booth, we’ll be demonstrating how Intel RSD composability works and run some benchmark applications to show off the performance of NVMe over Fabrics technology. If you’re going to the conference, come by the Intel booth to kick the tires and see what a difference Intel storage technology and NVMe over Fabrics can make.

Also check out Supermicro’s brand new RSD Solution Center where you can evaluate the flexibility, performance, efficiency and manageability of Supermicro’s RSD systems. The center supports both on-site visits and remote access, allowing you to dynamically configure hardware resources for specific workloads using the Supermicro RSD graphical user interface (GUI) and command line interface (CLI). To start experimenting with your own workloads, contact Supermicro at: total_solutions@supermicro.com.

Intel RSD compatible products are available now from major vendors including Dell EMC*, Ericsson*, HPE*, Huawei*, Inspur*, QCT*, Radisys*, Supermicro*, Wiwynn* and others. For more information on Intel RSD, visit https://www.intel.com/intelrsd.

Be sure to follow me on Twitter to learn more, find me @UlgenFigen.

*Other names and brands may be claimed as the property of others.
Published on Categories Cloud ComputingTags , , , , ,
Figen Ulgen

About Figen Ulgen

Figen Ülgen, Ph.D, General Manager, Rack Scale Design Group, Intel Corporation. Figen Ülgen is General Manager of the Rack Scale Design Group within the Datacenter Group at Intel Corporation. Ülgen oversees solutions roadmap, solutions enablement as well as operations, support and ecosystem and partnership strategy for Rack Scale Design. Ülgen joined Intel in 2014, where she led as General Manager for HPC Software and Cloud developing and promoting HPC Platform Software technologies for the HPC ecosystem. She founded the award winning OpenHPC Community under Linux Foundation with a group of ecosystem partners to democratize HPC. During her time with the HPC Software and Cloud, the team shipped Intel® HPC Orchestrator, a fully supported production quality software product in less than 18 months—concept to General Availability. Prior to that, she was a Senior Director with Microsoft for over a decade in various leadership roles in Data Science, Azure Public Cloud and HPC teams. Her 20 year industry experience includes technology innovation, competitive strategy, software product planning and management, as well as driving change across large organizations. Before Microsoft, Ülgen worked at McKinsey & Co., Motorola and at Justsystem Corporation in Japan. With a passion for serving underprivileged youth and women in technology field, Ülgen sits on the Saturday Academy Board of Directors and is regularly invited to speak on behalf of women in technology. Ülgen earned a Bachelor’s Degree in Computer Engineering, a Master’s Degree in Expert Systems under a Fulbright Scholarship and a Ph.D in Machine Learning under the Japanese Government Ministry of Education Scholarship. She holds two U.S. patents.