Wrapup on Backup from VMworld 2015

data-protection-signIt's taken me a while to catch up after VMworld, but I did want to share some observations on what was interesting at the show from a data protection perspective. Thankfully, nearly gone are the days of asking “How do I get a good backup of my VMs?

With the prevalent usage of VMware’s vSphere APIs for Data Protection (VADP), most of the currently shipping backup products can now at least get a good backup. Of course, there are unfortunately far too many folks running older versions of their backup applications, under the pretense of wanting a bug-free mature version of their backup platform. But those folks that are running the older versions aren’t able to take advantage of the newer VADP APIs and therefore are still struggling to get a good backup. The rest of us, who are past that, were really talking about two things at VMworld 2015: Agility & Hybrid

From an agility standpoint, with the presumption that a good backup can be made, folks are looking at:

  • How fast can I get my VMs running?
  • How granular can my recovery be?
  • What options do I have for how that recovery will occur (e.g., where, by whom)?

I could spend another blog (or an hour or a day) on options and ramifications for the question, “How fast can I get my VM running?” -- and discuss the idea of when any backup is not enough versus when should snapshots or replicas be used for recovery. Moreover, where does the line blur between ‘rapid or instant recovery’ as a reactive measure and the proactive stance of ‘availability?’ That is a question being asked by many and whose innovations are being led by only a few.

The latter two questions combine to address how flexible your recovery is, including:

  • What type of restoration will you use -- backups, snaps, replicas
  • From which medium -- disk, tape, cloud
  • Being invoked by whom -- backup admins, vAdmins, or IT operations
  • With what type of scale -- file, message, folder, volume, VM, or multi-VM app
  • To what kind of destination -- same host, different host, in the clouds

And on the last point of ‘where’ … everyone wanted hybrid scenarios including 'the cloud,' but which cloud?

  • Which cloud will your production VMs be in? – your on-premises private cloud, vCloud Air, a hybrid infrastructure including public-cloud hosted storage and compute, etc.
  • Which cloud will your protection be delivered through? – BaaS, DRaaS, or maybe your own servers protecting cloud-based production servers (i.e., ‘hybrid in reverse’)

Key Takeaways on Data Protection from VMworld 

I wish that I had some great revelations on which paths or products looked most interesting, but the most impactful impression that I walked away from VMworld with was that many folks want to blur the marketing lines and their actual product features, including which features are shipping, planned, or simply aspirational … with even the buzzwords seemingly adapting to each customer request.

  • The good news is that most of us are asking the right questions, because data protection has always been, is, and should be about ‘recovery’ instead of ‘backup.’
  • The bad news is that everyone is saying that they do “that” with quite a bit of variation in the details between the leaders and the poseurs.

For something a little more definitive and cheery, check out my other blog post on some of the great marketing efforts that made VMworld 2015 effective.

 

decentralization of data protection

Topics: Data Protection VMworld