Okay, some more riffing on the subject of VPLEX; just idle speculation and I expect either no comment or complete denial from EMC.
In VPLEX, do we see the future of EMC storage? Or more accurately, do we see the future of Symmetrix? Is it the beginning of the end for Symmetrix and more importantly, Enginuity. The final break from the past?
It’s certainly an interesting thought; firstly despite regular denials from EMC; bushfires regularly break out between the Clariion and the Symmetrix camps. Should all the development efforts be focussed on Flare or Enginuity? Some people might say that at times, there have been more than bushfires. This is to be expected, when development teams are competing for finite resource, this sort of thing happens. But this never really gets resolved, especially when you have two successful products like Symm and Clariion.
Perhaps what is needed is a break from the past? Could VPLEX be this break? Well certainly from what has been announced so far, probably not. But if we look at the various EMC blogs where it is suggested that they might add things like snaps, clones etc; this seems a possible way forward.
VPLEX as an array controller may well be more interesting long-term as opposed to this ‘new category of storage device’; a device which has a much looser coupling with it’s back-end disk than the current range of storage arrays. There are reasons why EMC might want to do this or at least customers might be interested in such a device.
I posit that array controllers are actually changing quicker than the back-end disk these days; sure, disks are getting bigger but this is not entirely beneficial to especially the Enterprise storage market but if I want to take advantage of the latest greatest features of EMC’s latest greatest array; I have to rip out both the array controllers and the back-end disk.
What if I didn’t have to do that any more? What if I could upgrade the array controller completely separately from the back-end disk? What if this was a completely non-disruptive upgrade? What if in order to do a migration, I didn’t have to temporarily have twice as much disk on the floor as I do in normal operation?
Perhaps at that point EMC have actually built a truly modular storage array? And what if EMC can finally head towards a unified code-base for block storage? Yes, they might have to live with three code-bases for a period of time but it might actually be a worthwhile investment for them or perhaps they are happy to continue with a multitude of code-bases.
Just thinking what I might do in their situation…