Why NVMe Is Better When Shared

Data Center

The NVMe (Non-Volatile Memory Host Controller Interface Specification) standard is being adopted at a rapid pace, with many enterprises already designing all-NVMe data centers as part of new shared-nothing storage architectures for web-scale applications. A recently awarded patent highlights the innovation taking place around NVMe-enabled storage, particularly if your IT network is running real-time analytics, visualization, simulation, machine learning, or other hyper-scale computing applications.

Excelero, a San Jose based technology innovator offering software-defined block storage that just launched from stealth in March, was recently assigned a US patent for underlying technology used to build storage architectures at data center scale using standard servers and state-of-the-art flash storage. The new patent covers Excelero’s technique to pool NVMe by combining two industry standards, remote direct memory access (RDMA) and storage-oriented memory mapped input/output (MMIO), which dramatically reduces CPU use, removing the target-side CPU from the data path, lowers latency and enables superior performance compared to traditional storage architectures. Excelero makes this technology available in its flagship NVMesh™, a software-only server SAN solution.

An estimated 70 – 80% of enterprises will have adopted a web-scale IT architectural approach by 2027– architectures that demand new levels of flexibility and efficiency. This shift is driven by new hyper-scale applications, which need to capture and store more data than ever before, and process data at real-time. To meet these new storage requirements, enterprises are turning to NVMe storage, yet find NVMe utilization rates averaging around 25%, as NVMe by itself can only really be consumed locally, in-server.  As a result, customers often end up with stranded capacity at various locations throughout the enterprise.

The new technology developed by Excelero and incorporated in NVMesh virtualizes the NVMe devices and unifies the capacity into a single pool of scale-out block storage, as if it were local Flash.  The approach makes data locality irrelevant and is a breakthrough in enabling local NVMe latency and speeds over the network, using standard hardware. It is what enables optimal NVMe storage utilization (up to 85%) and greater flexibility. It also allows the convenience of centralized storage while avoiding proprietary hardware lock-in and reducing the overall storage TCO. This allows enterprises to deploy NVMe devices at data center scale with maximum ROI.

A 100% software-only solution, Excelero’s NVMesh also shifts data services from centralized CPU to complete client-side distribution, scaling performance linearly at near 100% efficiency. Its flexibility allows for both physically converged or disaggregated deployments to create a virtual, distributed non-volatile array.

Amazon, Facebook, and Google have redefined IT for web-scale applications, leveraging standard servers and shared-nothing architectures to ensure maximum operational efficiency, flexibility, and reliability. As new application workloads – cloud, mobile, IoT, machine learning, and real-time analytics – drive the need for faster and more scalable storage, Excelero aims to enable enterprises to optimize their infrastructures in the same way as these tech giants. The patent represents the first US patent officially awarded to Excelero, with 9 additional US patents pending.

Share on facebook
Share on twitter
Share on linkedin
Share on email
Tom Leyden

Tom Leyden

Tom Leyden is vice president of corporate marketing at Excelero, a disruptor in software-defined block storage inspired by the tech giants’ shared-nothing architectures for web-scale applications. He has over 20 years of expertise in developing and marketing next-generation enterprise storage technologies.