United States
Site Map Contacts Hitachi Global Community
Hu's Blog - Data Storage and Virtualization Thought Leader Hitachi - Inspire the Next

Hu Yoshida's Blog - Vice President | Chief Technology Officer

Home > Corporate > HDS Blogs > HDS Bloggers > Hu's Blog
Products, Solutions and more

Hu's Blog

Unleashing the Power of Hitachi Virtual Storage Platform to VAAI

by Hu Yoshida on Feb 8, 2011

Those of you who frequently visit the VMware HCL (Hardware Compatibility List) website for SVD or Storage Virtualization Devices will have noticed that the Hitachi VSP is listed for ESX 4.1 with a footnote. That footnote will have a very important impact on where, when, and how you can use VMware with existing storage systems.

I have invited Michael “Heff” Heffernan to post his perspective on this footnote. Michael has been spearheading our technical partnership with VMware and has been largely responsible for our integration with VMware features and functions. Before joining Hitachi Data Systems, Michael was a customer who had experience installing and tuning VMware, so he can also provide input from a user’s perspective. Please feel free to leave Heff a comment on this blog post and he will be happy to respond. – Hu

virtual-server-imageA little less than seven months ago, the cooperation between VMware and the ecosystem of storage partners changed the face of virtualized server environments with the introduction of VAAI (vStorage APIs for Array Integration for those of you not yet familiar with the acronym). By providing a way to offload key processing I/O from the ESX host environment onto a storage array, VAAI not only gave organizations a means to scale their virtualized environments easily but also removed a lot of complexities the VMware and storage admins had when operating their VMware infrastructure on a daily basis.

However, as with really anything new – especially technology – there are some important considerations to think about when making any VMware infrastructure upgrades. The challenge now for customers is learning how to easily take advantage of VAAI with vSphere 4.1, and they’re learning that all they require is for their storage array to support this in the microcode.

Sounds quite simple, right?

Not exactly. Up until now (seven months after VMware released VAAI support with vSphere 4.1), there are still storage arrays in the market that simply will never support VAAI, or even current arrays where  microcode will never be available to back-port to these older arrays, because they just can’t.

To get this functionality, customers are finding they will either have to “rip and replace” their existing infrastructure for a new VAAI-enabled array and go through the pain to migrate 100s to 1000s of VMs onto the new platform or just sit around waiting for their vendor to support VAAI on their future roadmap. Either way is not only going to cost the customers both from an OPEX & CAPEX, it also will introduce greater risk of downtime to the business. So, the problem that is now faced by many customers is how can they take advantage of the power of VAAI while protecting their ROI?

Here’s how …

Today, Hitachi announced certification of VAAI on Hitachi Virtual Storage Platform and — importantly — with external storage. What this means is that we’re able to extend VAAI benefits not only to ESX hosts connected directly to  our VSP but also to the 100+ storage systems that can be virtualized behind it!

Now you’re thinking – very innovative? Well, we think it is…

This is really something special for the entire VMware infrastructure community; in fact, HDS is the only storage vendor to support all three of the VAAI “primitives” on a virtualized storage platform.

If you really want to appreciate the value of VAAI, be sure to read Hu’s blog post: “It Rolls Downhill”. It does a great job explaining how Atomic Test and Set command eliminated all the contention that the array faced when multiple ESX hosts accessed the same volume simultaneously. This has allowed customers to leverage the true benefit of the VMFS clustered file system with up to 85% performance improvement and increase the VM density on a single VMFS volume. Based on my interaction with customers, many are embracing this change and experiencing massive operational benefits with their VMware infrastructure especially for arrays that have been purposely built for this type of integration.

Hitachi’s Intelligent Formula

Hitachi has now developed the true formula to support any scalable data center transformation of virtual servers with the ability to provide choice and flexibility in conjunction with protecting customers’ existing storage assets. I’ll write more on this topic in my next post.  Stay tuned!

– Michael Heffernan, global product manager, Server Virtualization, Hitachi Data Systems

Michael Heffernan is the global product manager for Server Virtualization at Hitachi Data Systems. In this position, Michael ensures that the core technologies of Hitachi storage and servers are aligned with hypervisors in order to create integrated technology solutions that will benefit Hitachi customers. Previous to this role, Michael worked in Asia Pacific as director of Service Delivery at Hitachi Data Systems, where he leveraged the company’s delivery methodology and storage service tools to deliver a superior solutions service experience for customers.

Related Posts Plugin for WordPress, Blogger...

Comments (13 )

Greg Knieriemen on 08 Feb 2011 at 9:32 am

This is huge – and a significant announcement for the majority of VMware shops that use multiple storage vendors.

Fabio Rapposelli on 08 Feb 2011 at 9:53 am

This looks great, definitely a huge improvement for HDS in the High End space, but a question comes to mind, will it work between different virtualized arrays? e.g. between a virtualized Clariion and a virtualized AMS?

Now all you need is to add a scaled-down external virtualization capability to your AMS midrange line (just like IBM Storwize 7000V) and you’ll have a killer storage lineup.

Fabio

Rick Vanover on 08 Feb 2011 at 10:13 am

Yes, this is good stuff.

I am not sure if any other storage platform has every model supporting VAAI.

Rick Vanover on 08 Feb 2011 at 10:41 am

I will also agree with Fabio that I’d love to see a modular for the smaller environments (Sub 10 TB).

the storage anarchist on 08 Feb 2011 at 3:34 pm

Just a point of fact: in the 2nd half of 2010, both Symmetrix VMAX and CLARiiON CX began shipping a software update that added VAAI support for the installed base of both products, and the new VNX Series supports VAAI right out of the gate.

TimC on 09 Feb 2011 at 1:43 am

@Hu: Quick fact check. It would seem to me NetApp was first with a virtualized array supporting VAAI. They supported VAAI on ONTAP 8.0.1 with their v-series array. It went GA several weeks ago.

@anarchist: He pretty clearly stated the first virtualized storage array. Neither VMAX or Clariion virtualize. And “second half”? Is that what you call December 15th? I’d call that “barely 2010″.

[...] Storage Platform both for the product itself and when used with external storage.  As stated in yesterday’s accompanying blog post, we are the only storage vendor to support all three of the VAAI “primitives” on a virtualized [...]

Heff on 09 Feb 2011 at 5:28 pm

Thanks all for your feedback and comments.

@Fabio – Excellent suggestion and feedback. We have had a lot of success with our USPVM so this could be a solid approach for the future.

@Rick – Currently when reviewing vmwares com HCL that’s correct not all storage devices are supported as yet. Therefore the VSP is a great option for customers who require VAAI and do not want to wait or rip and replace theses storage arrays. HDS has been assisting customers for the last 8 years for virtualizing assets in there data center to obtain many benefits like central point of management, tiered storage management & migrations – this is just another example of the innovation of this technology for the use case of VAAI for these externalized storage devices.

@Barry VSP currently supports VAAI all 3 x primitives for external storage devices – this means VSP can virtualized 100+ arrays that may not have VAAI supported – A great solution for customers who want extend the life of their asset and save on future costs without having to rip and replace hardware.

@TimC It appears that the vmware HCL is not up to date. I would like to understand if V-series 801(CIFS & NFS only) supports VAAI on EXTERNAL storage ?

Cheers,
Heff

Pete Gerr on 09 Feb 2011 at 9:59 pm

@TimC – Not sure if you’re a NetApp employee or not but I admit it’s difficult for anyone to keep the many differences between 7-mode, C-mode, GX, 7G, etc, etc, etc straight; I know I was the ONTAP PM for quite some time.

But, back to the real questions and topic:

The main point is that the VSP is the first and only (as far as we can tell from currently published material on currently shipping products) certified SVD to support all 3 VAAI primitives AND to extend this support for all 3 VAAI primitives to any of the 100+ 3rd party arrays supported as external, virtualized storage behind the VSP.

I’ve posted the following questions on Twitter and still haven’t seen a definitive answer, so perhaps you can clarify:

1. The URL you provide above does show V-series and that the VAAI primitives are supported for the V-series but I believe this is only for internal disk inside the V-series, correct? This is not stating the V-series is certified as an SVD…

2. If #1 is correct, this means that any 3rd party external storage behind a V-series does not inherit all 3 VAAI primitives.

3. Correct me if I’m wrong, but the SVD designation means that a given system is certified as a “Storage Virtualization Device”; the current list of certified SVDs is here – http://bit.ly/fBv8sl

Please note that keyword “SVD” is included within my search but was not included within your search link you provide above. V-series is on Page 2 of my search list, but shows that the 3 VAAI primitives are not supported on the V-series as an SVD.

4. Stated a different way – VSP can push all 3 VAAI primitives to external storage arrays that have been virtualized behind a VSP.

Real-world example: Roll a VSP in front of a DMX or CX (neither of which support all 3 VAAI primitives as an SVD). Because HDS virtualizes within the storage controller, we enable the DMX, CX, etc, etc, etc to inherit functionality they never had and, likely, never will. So the question I have is with a V-series running 8.0.1 7-mode without internal disk and with say, a CX behind it, does the CX inherit all 3 VAAI primitives? My belief is no, but that WAFL/ONTAP remaps the LUNs on the CX making it impossible for the V-series to push all 3 VAAI primitives to the CX.

For customers, this is great – investment protection with the storage they already own, extends the functionality of that storage with important features that Hu and Heff outline above.

Looking forward to your response and thanks for all this lively discussion!

Pete

TimC on 09 Feb 2011 at 7:58 pm

CIFS and NFS only? What are you talking about? Ontap 8.0.1 supports FCP/FCOE/iSCSI/CIFS/NFS/HTTP/etc/etc/etc. I assume when you say “CIFS and NFS only”, you’re talking about cluster-mode. This is probably a good place to start if you’re unsure what your competitors are bringing to the table:
http://blog.bhagat.com/wordpress/2010/11/netapp-ontap-8-0-1-7-mode-new-features/

As for the HCL, the v-series is there and supported:
http://www.vmware.com/resources/compatibility/search.php?action=search&deviceCategory=san&productId=1&advancedORbasic=advanced&maxDisplayRows=50&key=&release=24&datePosted=-1&partnerId=64&arrayTypeId=-1&firmware=-1&isSVA=0&rorre=0

TimC on 10 Feb 2011 at 12:34 am

No, I’m not a NetApp employee. As for the confusion, I’ll leave it at this. There are two versions of ONTAP, 7g and GX aka cluster-mode. The nomenclature carries over to the 8.x code-lin. I’ve yet to meet a customer who is confused by the fact there are two versions of ONTAP. Just like I’ve yet to meet a customer who has difficulty understanding the difference between AMS and USP/VSP.

1. It’s supported with VAAI as an SVD, per the support matrix.
2. No, it isn’t “only for internal disk”.
3. http://bit.ly/eXSfOu – SVD.
4. They both can.

Why would you think a LUN behind a v-series couldn’t do VAAI? Having third-party disk behind a v-series is no different than having native disk. The only thing that changes is that ONTAP skips doing RAID because it’s handled by the back-end disk. I’m not sure what you mean by “remapping”, but the LUNs are not simply passed through a v-series.

The only reason I can assume you think that VAAI isn’t supported is because there is a footnote that you need a plugin for VAAI to function with the HDS array, and not NetApp. There’s no footnote on v-series for the same reason there’s no footnote on FAS: because there is no special plugin required.

As for hearing it directly from a NetApp employee:
http://blogs.netapp.com/virtualstorageguy/2010/07/vsphere-41-and-vstorage-apis-for-array-integration-vaai.html

Peter Learmonth confirms it in the comments on that blog.

The VSP is an awesome box. I wouldn’t get so caught up on whether or not you’re first, because quite frankly, customers don’t care.

[...] quite a bit of dialogue in blog circles and on Twitter this week about our recent announcement of VAAI certification on Hitachi VSP with external storage.  Predictably, much of this came from competitors on our statement that [...]

[...] launch date and in February of this year, they announced VAAI support with their VSP (see also Hu Yoshida’s writing on the [...]

Hu Yoshida - Storage Virtualization Thought LeaderMust-read IT Blog

Hu Yoshida
Vice President and Chief Technology Officer

Connect with Us

     

Switch to our mobile site