VMware Virtual SAN from 5.5 to 6.2
Remember in Feb 2013, VMware acquire Virsto a block storage optimizing software company and liquidise in it Jan 2014. This was nice piece of technology which took everyone by shock that it's no longer offered as a product. How good can it be to serialize data on a block storage system when it was written in random.
With VSAN 6.0 release, a part of Virsto was present in its file system and with it's native snapshot. Yes it is not using the traditional snapshot offered through vSphere which is still done on any other storage. Now it is more applicable for Business critical apps as it also offer All-Flash VSAN model.
In VSAN 6.1 release, it allows stretch VSAN cluster. It also extend vSphere Replication with VSAN to have a RPO time of 5 minutes. It also comes with support of Oracle RAC and WSFC. Not forgetting SMB customers, it also comes with a ROBO offering with a two nodes setup. Deduplication was added as a tech preview in VSAN 6.1.
What was missing in the previous versions which many competitors who offer their SDS solution would mention would be deduplication, code erasure, compression and encryption.
Other small features not generally mentioned is Client Cache where you can allocate up to 1GB of memory per host for read ahead cache. Data Locality will be applicable here. For Horizon View, this will work inline with Content-Based Read Cache (CBRC). I am going to see hell lots of improvement for Horizon on VSAN in areas of "storms".
Another small advanced option not enabled by default is Sparse Swap. This allows memory swap file to be thin provision instead of the traditional upfront provisioned. E.g. VM created with 4GB memory will have a memory swap file of 4GB place on the disk. Imagine how much storage is required if every host has 512GB of memory utilized by 50% of memory by VMs in a 64 hosts cluster? With that advanced setting enabled, all swap file will be thin provisioned and allocated when consumed.
With the announcement made, the binary isn't ready till March 2016. So let's wait and see.