VMware vSphere and the vStorage API

During my trip to Palo Alto for the VMware vSphere Launch I met with some NetApp engineers at their Executive Briefing Center in Sunnyvale.  This was perhaps one of my favorite meetings while in the Bay Area that week.

There are going to be a lot of exciting things coming with the GA release of vSphere, one of the biggest in my opinion will be with vStorage API’s, which will allow vSphere to offload tasks to the storage subsystem.

I was told by NetApp that in vSphere when you initiate a Clone of a Virtual Machine or Deploy from Template, this process will be sent to your FAS system to process, rather than it happening at the ESX host level as it traditionally was done.

Other things that will be sent to the storage system include De-Duplication (if your storage supports it), commands sent by VMware Data Recovery and also automated provisioning in VMware View can be offloaded.  For VMware View, NetApp provided the example of their Rapid Cloning Utility and how it already integrates with the VI Client – the new release will be a lot more streamlined since it will use vStorage API calls.

So, what does all of this mean?  It shows that VMware is working close with the hardware vendors to truly build a Virtual Eco-System, and it shows a lot of exciting new potential that vSphere will bring. Pushing storage related processes back to the storage subsystem will not only increase the speed of these transactions, but it will also relieve a lot of CPU and I/O from the ESX hosts.

Created on May 5, 2009 by Rick Scherer

Posted under NetApp, Storage, vSphere.

This blog has 9,714 views.

Tags: , , , ,

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

4 Comments so far

  1. Scott Vessey
    3:17 pm on May 5th, 2009

    I agree that this is a great new capability in vSphere 4, there are many operations that an admin performs that can take advantage of this :-)

  2. Vitor Andrade
    6:19 pm on May 5th, 2009

    Really a great improvement.
    Setting free resources from CPU cycles, intensive I/O can be used to other purposes, specific cases of use and so on.

  3. Peter E
    1:38 pm on November 7th, 2009

    The vstorage API doesn’t send “deduplication” to a storage system. It enables block-level incremental backup (BLIB) which makes for a much more efficient backup. Deduplication would be the elimination of redundant blocks or files within AND across multiple VMs (and ideally phsyical machines as well) in the backup set(s).

  4. Rick Scherer
    1:54 pm on November 7th, 2009

    What I meant was the ability to enable data deduplication via the vSphere client…for example; enabling NetApp A-SIS on a specific volume, etc.


Leave a Comment

Name (required)

Email (required)



More Blog Post