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

CIOs Understand Storage Virtualization

by Hu Yoshida on Apr 17, 2010

I was recently asked by an analyst firm to review a survey that they intended to send to end users to evaluate the extent of virtualization in the data center. While most of the survey asked very relevant questions, when it came to storage virtualization, it was all over the map, primarily because there was no clear definition of storage virtualization. The list of storage virtualization vendors seemed to include every storage vendor in the world. Some did not do any form of virtualization other than RAID, some were associated with server virtualization, and the majority of them did not do any storage virtualization outside of their own box. Only a few of them fit the SNIA definition of storage virtualization which I have referred to many times. My recommendation to them was to define storage virtualization up front or else the results of the survey would be meaningless. The industry has made a mockery of the term “storage virtualization.

That is why I was really pleased to see a panel of customer CIO’s bring this back into focus at SNW this past week. On Tuesday there was a panel of CIO’s discussing “Promising Technologies That Matter Right Now”. Dave Raffo of SearchStorage.com filed a report of this session on April 14.

The panelists were Anthony Abbattista, vice president of technology AllState Insurance Company, Martin Gomberg, CIO, A&E Television Networks, and Jeff Kuback, CIO, Kroll. The key technology that they all identified was storage virtualization.

Gomberg of A&E said that the technology that held the most promise was storage virtualization. It was an enabler for consolidation of the data center. His managed storage was already in the 45 PB range!

Kuback of Kroll talked about the data mobility that was enabled by storage virtualization. Specifically he was planning to migrate older heterogeneous storage systems to new systems in a matter of months rather a year without virtualization. He would be using the Hitachi Data systems USP-V for this migration project.

Abbattista of AllState saw storage virtualization as an integration layer that would enable him to buy any storage and plug them into a common management platform, and move a tier of storage like interchangeable parts.

After seeing so many analysts and vendors talking about storage virtualization in so many ways, I was really pleased to see that these CIO’s clearly understood what storage virtualization is all about. It is not about virtualizing the capacity inside one’s proprietary storage array. It is not about proprietary file systems with virtual mount points. It is not about virtualization that is tied to a proprietary operating system or host based software.

It is about consolidating heterogeneous storage systems into a common pool of storage, with common management, common protection, and common search that can scale to PB. Storage virtualization should be independent of the upper layers in the stack and give you a choice in selecting other vendor’s storage arrays and integrating them under a common management layer with the latest enterprise features like data migration, dynamic provisioning, and dynamic tiering.

These CIO’s clearly understood what storage virtualization is and the benefits of virtualizing external storage arrays from different vendors.

Related Posts Plugin for WordPress, Blogger...

Comments (5 )

Josh Krischer on 19 Apr 2010 at 9:53 am

Hu,
In my opinion storage virtualization is one of the most misunderstood techniques.
In my presentation on Storage Virtualization I listed nine techniques however many people see #3 as Storage Virtualization per se:

1. Redundant array of independent disks (RAID)
2. Volume management
3. SAN Virtualization (block)
4. Virtual File systems
5. Logical Partitioning
6. SSDs
7. Virtual tape
8. Virtual ports
9. Virtual SANs

Barry Whyte on 20 Apr 2010 at 10:07 am

Hu, by these definitions, surely another key factor should be the ability to enable ALL disk purchases to be heterogeneous if required, and the choice of virtualization platform shouldn’t necessitate the purchase of disk and a enterprise controller from the same vendor too?

You’ve always tried to say SVC isn’t storage virtualization, but in every one of those client definitions and infact your own definition, I’d say SVC fits that description better than USP-V ?

Barry

Hu Yoshida on 23 Apr 2010 at 3:09 pm

Barry, the first part of the SNIA definition for storage virtualization calls for independence of the storage from networks as well as applications.The SVC is part of the SAN. I believe the developers of SVC call it a SAN Volume Controller
You can buy a USP VM without buying any storage from HDS.
So we clearly meet all definitions of storage virtualization.

Hu Yoshida on 23 Apr 2010 at 3:36 pm

Josh, Cache controllers precede RAID and were probably the first form of virtualization. You and I can remember writing control programs directly to spinning disk and processing Count Key Data commands during the gap time between data blocks.As the density of the disks increased, the gaps consumed more and more data space.Ibelieve we can thank Moshe Yanae for the development of cache controllers that enabled us to process this in cache and use lower cost Fixed Block Architecture disk.

GregR on 06 May 2010 at 8:50 am

Hu, by this definition, wouldn’t you need to include just about every cloud storage vendor in the bucket of Storage Virtualization?

After all, you can buy those platforms without any local storage as well (well, most of them anyway), and use existing SAN storage for local cache purposes.

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