Lately, one of the question that keep coming during any Software Define Data Center architecture session with my customers: “Is VMware VSAN ready for production environment?”. The short answer is: it depends, for a longer version explanation please keep reading.
Before I got into details of the answer to the question if VMware Virtual SAN is production ready or not, let’s give a bit of background of why this question keep raising up. Initially when VMware released the VSAN beta mid last year, the focus use cases was VDI, ROBO, & some Dev/Test. That is quite understandable for a beta released product, but as a lot of the literature of the beta days are still out there, many are confusing what the beta was able to deliver in comparison to what the GA version offer on the other hand.
Further, few VMworld sessions this year when the product went GA, did not update their use cases or shied from having a production use case as a part of the use cases showing on the slide. Below is an example of the slides showing VMware VSAN use cases while leaving production use cases out:
Though not every presentation at VMworld left the production use case of their slides, but if you only went to one VSAN session and that session left it out you left with the ideat it was not production ready. Below is another couple slides of VMworld US, where the production use cases showed up:
Again the two slides above although they have included some production uses cases for VMware Virtual SAN they have always limited it to Tier 2/Tier 3 which has raised another question if its ready for Tier 1 production.
Defining if VMware VSAN is ready to handle your production environment and which particular tier of your production environment it can handle is highly dependent on your production environment requirements. For example, by default VSAN support a minimum of 15 minutes RPO, so if any of your production environments require a lower RPO, then VSAN might not be the right solution for that part of your production environment. Who ever put the above slides have assumed that Production Tier 2/ Tier 3 does not require such a low RPO at most customers, which make VMware VSAN a good candidate for such a workload. Further, they made the assumption that Production Tier 1 usually requires RPO lower than 15 minutes and in many cases require a synchronous replication which is not yet supported by VSAN, but being worked on for future releases.
Your definition of production environment might be totally different than the assumptions above. For example, there is many small businesses whose production Tier 1 have an RPO of 2 hours & upward and for those VMware Virtual SAN will fit their production Tier 1 environment. On the other hand, there is large enterprises or enterprises where the availability of their services is so critical for their continuous operation that even Production Tier 2 & sometime even Tier 3 require an RPO less than 15 minutes then VSAN is not the solution for their production environment today, but its getting there in the near future.
As far as reliability, performance, & scalability VMware Virtual SAN is definitely production ready and in many cases can beat traditional storage on these measures, but as with any storage purchase you are planning on, you will need to ensure proper sizing is taking into consideration and the size you are building can fit your requirements. VMware has already released sizing tools to help with that, and you can reach out to your VMware SE to help you with properly sizing it.
Hope this help cleaning up this misconception, and looking for others feedback and experiences with VMware VSAN.