EMC today announced the their latest entry into the Software Defined Storage (SDS) market, VIPR. They’ve coined it the “World’s first Software Defined Storage Platform” (http://www.emc.com/about/news/press/2013/20130506-03.htm). I have to say, I am a little put off by this initial push and need to be first when they are clearly not. I could list a few that have claimed to be a SDS platform first, DataCore, Nexenta, and when looking at some of the capabilities, I think IBM beat them out with the SVC Director.

Here is the summary of the feature set I see so far with VIPR along with some who may give some of the same capabilities. This is, by no means a complete listing I am sure, and since I don’t work for EMC, I can’t guarantee that all of these features will make the final release when it ships.

  • Storage Abstraction – The ability to discover existing storage arrays and their capabilities, then carve out new Virtual Storage Arrays.  I am not clear yet on this if VIPR uses the existing array’s capabilities for things like snapshots and replication or if VIPR will provide an extra layer of advanced feature set on top of existing arrays. My best guess so far is that it will do both. It should be noted that the initial release will likely only discover and present EMC and NetApp arrays.
    • Who Else has this capability:
      • DataCore – They provide a centralized management for multiple arrays when presented to their OS, but they do not use the capabilities of underlying arrays
      • Nexenta – Provides unified storage to hosts when presented with any SAS, NFS, iSCSI, or FC LUNs, however all LUNs are treated as raw disks so, there is no underlying array management or capabilities
      • IBM SVC – This attaches to the existing SAN Network and presents data back outbound. It does not require data to be transitioned to their own filesystem, which is contrary to the other solutions.
  • DataCenter Visibility – End-to-End visibility into the storage environment, which provides QoS reports, metering and a multi-tenant model. I will be looking for more detail on the multi-tenant still.
    • Who Else can does this?:
      • DataCore – Provides reports for the systems it manages, I’m unsure of the multi-tenant component
      • Nexenta – By using DTrace custom reports, which can be generated, or by using the built in reports, some of the knowledge could be garnered. There’s no information on the back end arrays. Multi-Tenancy is only done through virtual NexentaStor arrays
      • IBM SVC – Reporting is based on the XIV on which the latest interface is built
      • NetApp – Multitenancy is built with the vFilers.  Reporting is done for the entire Netapp suite but it does not front multiple arrays currently.

 

  • REST API for Management – This will also be EMC and NetApp only to start for what Chuck Hollis refers to as “Southbound adapters”.  The Northbound our front end seems to have what is expected with the addition of some emulation of object store
    • Who Else can do this:
      • Nexenta – API has been built into the product for a while I believe a few others have REST API already but this is not my area of expertise
  • Data Portability – I am calling it this, although it may not be the most accurate statement. The ability to have data stored on the system and originally be presented as NFS data and then when ready switch to present the same data as an object store, and then the next is FC presented.  This is something that I think has been missing for a while and while not specifically SDS a very cool concept.
    • Who Else can do this:
      • I can’t say that anyone can do this exactly.  The ability to wrap NFS on top of an object store or HDFS on top of object is rather unique.  Being able to present CIFS as NFS is possible with most of the unified players, although Nexenta allows you to present them both at the same time, which I believe is unique to them. Presenting all block level at the same time is also possible but switching between them is not, so I don’t see anyone else doing this yet.
  • Storage and Virtualization Automation – The ability to tie directly into the virtualization configuration for cloud architectures using vCO and having the storage emulation is unique to EMC. It does help that EMC has a unique tie into VMware. Future integration with OpenStack is also mentioned and since they can already present using Cinder, I would expect this to come soon.

 

Overall this is a great step in the evolution of SDS and I expect to see the other larger storage vendors follow suit. NetApp’s Vaughan Stewart has been talking SDS for a while and while EMC is clearly not the first SDS platform it is still a solid release and recognition that the need for hardware lock-in should not be the case. Seeing EMC and NetApp move to a model where the consumer chooses the hardware themselves would be a great next step but would erode the margin model significantly.

 

To me Software Defined Storage is still not fully realized or even partly complete.  When our applications start to drive storage, then we will see true SDS but until then… Thank you to EMC, DataCore, Nexenta, Netapp, and IBM for continuing to get closer and closer.

A few other fellow bloggers have also put out articles on VIPR today, take a look:

 

**Note: If you are not aware I do work for Nexenta currently but as usual attempt to be as objective as possible.