StorMagic® SvSAN is a hyperconverged infrastructure (HCI) solution that runs as a virtual machine on either Microsoft Hyper-V or VMware vSphere hypervisors. SvSAN has been shipping since 2009, during which time the company has built an installed base of over 1000 customers- many of which are running hundreds or even thousands of separate SvSAN implementations across the organizations’ locations.
SvSAN aggregates commodity, direct attached storage – HDDs, solid state devices (SSDs) and PCIe Flash devices – located in each storage node into a common storage pool and shares this capacity with VM workloads running on the hypervisor resident on each node. This hyperconverged configuration is most commonly used to provide a consolidated compute solution for edge computing environments like multiple branch offices, remote and branch locations and IoT deployments. The SvSAN cluster can also offer an alternative to traditional SAN-based storage arrays, sharing capacity with external server clients via iSCSI (servers need not run SvSAN).
Using low-cost, standard server hardware, SvSAN is deployed in a cluster of two storage nodes with data synchronously mirrored between the nodes. Using a local or remote witness agent running on another computer, StorMagic clusters can be expanded to thousands of locations, all controlled from a single management point. Both nodes can reside in the same location or at different locations within the limits of synchronous mirroring, with targets configured in an active-active mode.
SvSAN can be deployed over multiple remote sites through scripted deployments to automate the provisioning process. SvSAN can also be centrally managed through industry standard monitoring interfaces including SNMP v2 and v3, vCenter Event Forwarding and System Center Operations Manager. SvSAN provides several management tools, including a VMware vCenter plugin, a PowerShell CLI and MS Systems Center integration.
SvSAN supports “Predictive Storage Caching” with flash or DRAM, in both read and write-back modes. The software also supports read-ahead caching as part of the algorithm using system memory. This optional feature provides a server-side performance enhancement and can reduce the need for SSDs for performance and better leverage lower-cost hard disks.