Hyper-converged system startup Scale Computing has added hybrid flash-disk nodes to its existing all-disk nodes, boosting performance.
The existing HC1000, HC2100 and HC4100 nodes continue to be available. They are 1U boxes with a minimum three node cluster size, expandable to eight nodes and targeted at the SMB market.
Clusters can scale out to eight nodes, possibly more if you contact Scale. The systems run the KVM hypervisor, avoiding any VMware tax.
HC2150 and HC4150 nodes add a flash tier, with HEAT (HyperCore Enhanced Automated Tiering) placing data automatically on flash or nearline disk. Users move a scroll bar to tune per-VM flash priority allocation, from no flash to virtually all flash and so modify the automated allocation.
The HyperCore OS will “stripe data across multiple physical storage devices in the cluster to aggregate capacity and performance.” It also has a backplane network to let “any node and any VM access any disk and is performance optimised to scale as nodes are added.”
New nodes in the blue columns.
The new nodes can be added to existing clusters with existing workloads using the available flash automatically. Find out more about Scale Computing's software here (14-page PDF).