Need rapid networked storage? Take a look at NVMe-over-TCP

Lightbits helped develop NVMe-over-TCP to provide shared NVMe storage capacity over existing TCP-based networks

In data storage, speed is more essential than ever. In fact, 90% of IT organizations surveyed by ESG recently reported they now need to move faster than three years ago, with 41% accelerating their response speed by more than 50%.

For a modern business, competitive success requires scalable and highly efficient IT operations, but the rapid growth of data often leads to complexity in IT infrastructures, which puts IT teams in a tricky situation.

IT teams must juggle modernizing the IT infrastructure from one side, while at the same time simplifying operational management activities enough to allow fast reaction when needed. That’s where high-performance NVMe flash storage comes in, with the ability to create pools of NVMe storage across the network.

On average, organizations expect their on-premise data environments to grow in capacity by 35% annually, and they are on track to double in capacity in less than three years.

Meanwhile, NVMe-based storage is increasingly being relied upon to help storage environments keep pace with the growing need for high-performance data access.

NVMe provides a higher level of performance and lower latency than traditional protocols such as SAS or SATA, and to extend the low-latency benefits of NVMe across the network, NVMe over fabrics (NVMe-oF) has emerged.

According to the ESG research, 28% of IT organizations already use NVMe-oF, and 42% are poised to deploy it in the next 12 months. Organizations often opt for one of several NVMe-oF options, such as Fibre Channel, or other remote direct memory access (RDMA) fabrics. The challenge is that these are complex, expensive and often require specialized equipment. This is where NVMe-over-TCP comes in.

NVMe-over-TCP transforms the datacentre

NVMe-over-TCP is an industry-standard with a well-established ecosystem. Lightbits Labs worked alongside Meta (Facebook), Intel, Cisco, Dell EMC, and Micron to develop the NVMe-over-TCP standard, which was ratified by the NVM Express consortium in late 2018 as a layer of link transport. But what makes this standard so special?

NVMe-over-TCP provides low latency as if flash storage were local on direct-attached storage in the datacentre. But it is done over a network, on standard Ethernet TCP/IP networking equipment, and provides much higher levels of IOPS on the same Ethernet/TCP networks.

This approach is cost-effective because the fact that it is based on Ethernet means it does not require any new hardware investment, which makes it particularly attractive for hybrid cloud deployments.

Also, by perfecting the distribution of storage resources, NVMe-over-TCP allows IT teams to allocate just what is needed for applications and virtual machines. This results in a reduction in power consumption because unused space is no longer allocated.

NVMe-over-TCP is expected to be a key technology to drive mainstream market adoption of NVMe due to the prevalence of TCP/IP and ease of deployment. It makes networked NVMe easy to deploy at scale while delivering performance similar to local flash. By leveraging TCP transport, the standard helps to simplify network infrastructure management while ensuring supplier interoperability that allows for multiple options and flexibility for storage buyers.

That supplier support was considerably widened last September when VMware announced support for NVMe-over-TCP in a vSphere update. Key advantages for users are simplicity of deployment, a standard migration path using VMware Storage vMotion, and use and maintenance that are easy and non-disruptive.

Protocols such as NVMe and NVMe-oF enable ultra-fast, low-latency flash storage, which means businesses can run distributed applications, including Kubernetes, and access data in near real-time.

NVMe-over-TCP offers a considerable bonus in terms of cost, speed, and ease of deployment for organizations that require simplification as well as modernization of IT architectures.

In VMware environments, in particular, the new standard will help lower latency to a level that meets the demands of business-critical workloads.

Originally published at: ComputerWeekly.com

Sagi Grimberg is the chief technology officer at Lightbits Labs.