Automated Provisioning of High-Performance NVMe Storage over Ethernet Network to Servers Running Any OS or Hypervisor
DriveScale, Inc., the leading innovator of elastic bare-metal cloud infrastructure, announced support for the programmable Mellanox BlueField® SmartNIC, an I/O processing unit (IPU) that can be used to offload storage and network processing from the host CPU. The combination of BlueField and the DriveScale Composable Platform enables automated, end-to-end NVMe-oF deployment and the ability to compose compute nodes and storage over a standard Ethernet network then expand, reduce or replace resources on-the-fly while application hosts can run any operating system or hypervisor.
As enterprises invest significantly in big data analytics and machine learning, and Tier-2 cloud providers compete for customers, the challenge they face is delivering cost-efficient, highly utilized scale-out physical infrastructure that adapts instantly to the needs of the application. “DriveScale gives IT the flexibility and scale of networked storage with the performance and cost-efficiencies of local drives,” said Denise Shiffman, Chief Product Officer, DriveScale, Inc. “DriveScale turns commodity physical infrastructure into programmable components that are easy to deploy and adapt and with Mellanox BlueField, DriveScale now supports Windows and VMware application hosts in addition to Linux and other operating systems or hypervisors for data-intensive applications and machine learning.”
“BlueField provides NVMe-oF offload and NVMe SNAP to offload, virtualize, isolate and accelerate specific storage tasks from the host,” said Rob Davis, VP Storage Technology, Mellanox. “Together, DriveScale and Mellanox logically present networked storage over 25G/100G Ethernet as a local NVMe drive that can be instantly sliced, expanded and replaced providing flash storage virtualization and elasticity for bare-metal clouds, big data analytics and machine learning use cases.”
DriveScale’s server agent runs on the SmartNIC, sets up the NVMe/RDMA or NVMe/TCP data fabric and enables the automated allocation of drives or slices of drives to the host application as if they are local drives in the server. Using DriveScale’s API or GUI, users can set parameters for the application’s infrastructure and the compute, storage and network resources are automatically connected and ready for application deployment. Additions or changes to these resources are implemented on-the-fly.
DriveScale instantly turns any data center into an elastic bare-metal cloud with on-demand instances of compute, GPU and storage, including native NVMe-over-Fabrics, to deliver the exact resources a workload needs, and to expand, reduce or replace resources on the fly. With DriveScale, high-performance Kubernetes clusters deploy in seconds for machine learning, advanced analytics and cloud-native applications at a fraction of the cost of the public cloud. Learn more at www.drivescale.com.