Jeff Miller attended EMC World with us this year and wrote up a quick post regarding his high level understanding of EMC’s new VPLEX.
I have finally grasped the concept of VPLEX after attending the 2 hour VPLEX hands on lab. VPLEX is the big buzz here at EMC World and to put it simply, it is like a virtual raid for your SAN that can go beyond the datacenter as well as also being storage agnostic. You simply carve up storage and present it to the VPLEX , the VPLEX then will claim the storage from the backend and it will present it to your servers. You will claim storage from one SAN, claim storage from a different san (it could be in the same data center or at another data center within synchronous distance) and then you raid the storage together. The servers behind the VPLEX just see this single virtualized storage unit that is being presented from multiple sources. Within VMware, when you Storage vMotion a server between storage and even sites, the data is already there so it seems as if it moves from site to site within seconds.
Some of the key points are:
- It is active/active, very resilient.
- It can support 8000 virtual volumes per VPLEX cluster
- the maximum LUN size tested by EMC was 32 TB
- They recommend using 8 gig fiber between VPLEX devices
- It can support a maximum of 5 ms latency across locations.
- The easiest migration path is through Storage vMotion (assuming you are fully virtual, like me)
Be sure to also check out Joel’s thoughts on the VPLEX.