With the announcement of ViPR 2.0 EMC is bringing some improvements and new features to EMC’s Software Defined Storage Software offering. The new and improved ViPR has major steps towards supporting a geo-distributed datacenter. For object storage your multi-site geo-replicated ViPR environment will now be presented as a single global namespace. The world of Active/Active writes is supported with strong consistency using a distributed erasure coding system, so you no longer have to choose between performance and redundancy. Objects are now cut into chucks and stored in containers, this type of hybrid encoding allows for a very low storage overhead. To allow for WAN optimized geo-disperse active/active traffic, ViPR looks for access patents and uses this information to accelerate the process. Object access is done via S3, Swift, or the Atmos API. Another change in the object space brings the Centera API as ViPR CAS. Today ViPR only supports the API, but a future release will allow ViPR to support Centera arrays.
ViPR has also expanded heterogeneous support for Hitachi arrays. Also now being offers is commodity platform support using a certification model. At launch the first supported platform is the HP SL4540 with others in the works. For non-supports arrays you can go the OpenStack route. ViPR now supports many other vendors’ arrays using this model until they are added natively.
Also announced is the ViPR block service that is being powered by ScaleIO. ScaleIO and ViPR together will allow you to have complete data service on commodity hardware, which leads us to Project Nile – but that’s another post.