Marvell Enables NVMe over Fabrics Storage Disaggregation for VMware Virtualized Data Centers

Marvell, NVMe, Fabrics Storage Disaggregation, VMware Virtualized Data Centers
Marvell Enables NVMe over Fabrics Storage Disaggregation for VMware Virtualized Data Centers

Marvell announced that its QLogic® Fibre Channel and FastLinQ® Ethernet adapter solutions enable NVMe™ over Fabrics (NVMe-oF™) technology in VMware vSphere 7.0. As data growth continues to skyrocket, data centers are grappling with increasing power consumption, complexity, and cost associated with the demand for greater storage bandwidth and capacity. The integration of Marvell’s Fibre Channel and Ethernet adapters into vSphere 7.0 allows low-latency, high-performance NVMe flash storage to be effectively shared, pooled and managed across a fabric resulting in cost-efficient enterprise and hybrid cloud data center scale-out architectures.

Computing Slides Away from the Cloud to the Edge

In enterprise and multi-tenant containerized data center environments, users of VMware vSphere 7.0 can benefit from Marvell technology by leveraging the latest innovations and advances in high-performance flash storage to be realized over proven Fibre Channel (FC-NVMe) and Ethernet RDMA (NVMe/RDMA) fabrics. Marvell’s QLogic 269x 16GFC and 2700 32GFC Series HBAs provide low latency, highly deterministic and concurrent FCP and FC-NVMe storage access. QLogic 2770 Series HBAs deliver 50% more IOPS, security backed by silicon root of trust and virtual machine telemetry powered by StorFusion™ VM-ID technology. This rich feature set enables performance-centric and latency-sensitive applications to securely scale on a virtualization-aware storage area network.

Marvell FastLinQ 41000 and 45000 Series Ethernet Adapters support NVMe/RDMA over both RoCEv2 and iWARP protocols. Starting with VMware vSphere 7.0, customers can leverage the NVMe/RoCEv2 capabilities of FastLinQ NICs while future-proofing their data centers for potential use cases of NVMe/iWARP and NVMe/TCP. FastLinQ Universal RDMA capability, combined with future support for NVMe over TCP, provides IT, managers, with the broadest choice of options to scale-out NVMe.

Today’s announcement also marks continued momentum in accelerating Marvell’s vision for end-to-end Ethernet storage, specifically the innovative Ethernet Bunch of Flash (EBOF) architecture for optimal disaggregated high-performance flash storage. By deploying vSphere 7.0, data center operators can use Marvell’s FastLinQ adapters to power Marvell’s EBOF architecture comprised of NVMe-oF Ethernet SSD converter controllers, SSD controllers and, Prestera® Ethernet switches. Edge-to-cloud data centers can significantly achieve greater performance and lower total cost of ownership (TCO) utilizing this scalable architecture.

How CIOs Can Meet Their Goals with a Limited IT Budget

“Data center modernization strategies are moving toward a shared infrastructure, disaggregation of resources, and dynamic provisioning of storage to provide the right amount of storage and performance based on workload and SLAs,” said Seamus Crehan, president and founder of Crehan Research Inc. “This announcement around NVMe over Fabrics for VMware vSphere can help solve a significant portion of the IT resource disaggregation challenge, enabling business efficiency and investment protection.”

“NVMe over Fabric technologies are unlocking the value of NVMe flash by delivering a variety of low latency, scalable and trusted fabrics,” said Vikram Karvat, vice president and general manager, Server Connectivity Business Unit at Marvell. “The introduction of Marvell’s QLogic FC-NVMe and FastLinQ NVMe/RoCEv2 technologies into industry-leading virtualization platforms from VMware enables end-users to leverage economies of scale when deploying NVMe without compromising the performance, reliability or manageability of their infrastructure.”

Previous articleAutomation Anywhere Announces Free Bots to Combat COVID-19 Business Disruptions
Next articleAccedian, Solace, and MobiledgeX Partner to Deliver Assured Performance for Next-Generation 5G and IoT Services