If you’re not familiar with ScaleIO, EMC defines it as a “software-only server-based SAN that converges storage and compute resources to form a single-layer, enterprise-grade storage product. You can also think of it as similar to VMware VSAN, but usable for things other than VMware.
Essentially, ScaleIO is software that runs on servers. It combines the servers into a pool and aggregates their direct-attached storage (DAS) into a single storage resource pool that can be shared by all the servers as if it were an actual physical SAN array.
This is a great way to follow the “bring the storage close to the compute resources” model. Since a ScaleIO cluster can start with only a few servers and scale up to thousands, it allows for the creation and handling of massively-parallel compute processing workflows.
ScaleIO can take advantage of any and all disks installed in the servers: HDDs, SSDs, and PCIe Flash cards.
Previously, servers in the same ScaleIO clusters needed to be more or less the same. EMC has added support of mixed physical and virtual servers, and supports asymmetric hardware configurations.
Version 1.30 makes ScaleIO more elastic. It’s now possible to dynamically add or remove resources from a ScaleIO cluster.
This version also adds data protection in the form of a two-copy mirror. This allows for support of both forward and backward rebuilds.
Lastly, ScaleIO 1.30 is supported by the ViPR Controller (more about that shortly).
ScaleIO 1.30 will be GA before the end of Q2.
