SimpliVity also announced that in an upgrade to its OmniStack software – to version 3.5 – it has added the ability to optimise placement of workloads across clusters of its hyper-converged products.
Microsoft Hyper-V support will exist via native Microsoft Server 2016 features and management tools, just as currently SimpliVity runs VMware ESXi natively in its products, with no traditional storage formats such as LUNs, to deal with.
SimpliVity was a pioneer – along with Nutanix – of so-called hyper-converged infrastructure, which combines compute, storage and networking in one box. This is a trend in part inspired by the modular hyperscale architectures pioneered by web giants Google, Facebook, et al.
It deploys its software as a virtual machine on VMware on commodity x86 server hardware, either as OmniCube appliances from SimpliVity, or as OmniStack software pre-installed on approved server hardware.
SimpliVity also supports the KVM hypervisor on limited release. This, however, is aimed at a different market segment to the VMware and planned Hyper-V deployments, said SimpliVity solutions architect director, Stuart Gilks.
“Our small and medium-sized customers are where VMware and Hyper-V dominate,” he said. “KVM has other strengths and appeals, especially to organisations with large engineering capabilities and service providers looking to provide support for OpenStack environments.”
More on hyper-converged storage
- Hyper-converged storage/compute, such as IBM’s modular xSeries, poses a threat to standalone SAN/NAS in the datacentre.
- Nutanix deal to put storage software on Lenovo servers with products aimed at datacentre use cases plus a 200-strong dedicated sales team could put pressure on its partnership with Dell.
Meanwhile, version 3.5 of OmniStack has introduced workload optimisation. For now, this will make use of VMware’s Distributed Resource Scheduler (DRS) and will allow customers to balance workloads and data according to CPU, memory, storage and data location across multiple instances of SimpliVity deployments.
Gilks said typical scenarios would be where new workloads are created or where customers want to move existing workloads around.
DRS automatically generates recommendations for workload placement. Users can set the feature to automatically carry out optimisation or to wait for administrators to decide on recommendations.
Currently, this will only work with VMware-based deployments, but, according to Gilks, the strategic aim is to provide equivalent capability where the hypervisors support it. “We will add functionality via a phased approach with more details when Windows Server 2016 is generally available,” he added.