Column

Why don't we have clustered FC block storage?

Arun Taneja
Of course, you could cluster a couple of NetApp boxes together and get failover, but they were still two boxes running two (or more) separate instances of file systems and applications. On the block side, you couldn't even do that. The two Clariions always acted and behaved like two separate boxes, unless you used dual path software, in which case you could get an application to failover from one storage box to another. That was all good for the 20th century.

Then, a couple of smart guys in Colorado decided to build storage systems using IP as transport, initially using a proprietary protocol (and eventually replacing that with

    Requires Free Membership to View

iSCSI). But, they didn't just do a box. They added clustering in a novel way so that one could start with two nodes and (at least, theoretically) add nodes ad infinitum. This clustering is for more than high availability. It is for scaling performance, capacity, I/O and yes, high availability. That single event changed block storage for ever. There are systems from LeftHand Networks, EqualLogic and one expected soon from Pivot3.

Storage blogs
Read what all of our expert bloggers have to say on data protection, storage networking and more. Click here.
The same thing has happened on the file side. A couple of guys in a garage, granted their garage was bigger than most mansions in Newport, R.I., decided to use a clustered file system to build a powerful NAS box that essentially did for files what the iSCSI box did for block storage. It gave you a "start small, pay as you grow" NAS that still looked and behaved like a single box from a management perspective. Think of this as file virtualization and HA [high availability] combined, if you will. This event also changed NAS forever. Now, we have powerful offerings from Isilon, Exanet, PolyServe and NetApp (GX), and several others in the market. And frankly, the impact of this event is still in its infancy. Ninety-five percent plus of the NAS out there is still the garden variety.

That brings me to the main thrust of my observation here: Why have we not seen any Fibre Channel-based block storage that follows the same philosophy as iSCSI or NFS/CIFS? Is it that there is an inherent technology problem making this happen on the Fibre Channel side? Why is the concept good for one and not for the other? Or, is it that the "two guys in a garage" are, in fact, working on building one, and it is simply a matter of time until we will see these systems emerge.

I have done some digging into this area and have basically come to the conclusion that the technology required exists, the concept makes infinite sense, but once again the big guys are too caught up with NIH (not invented here) and protection of their monolithic storage to really do anything about it. To make it work one has to start afresh, look at the set of new storage network chips, Astute Networks, iVivit, Aristos Logic, Aarohi (now Emulex), for instance, available in the market that, in my view, are required for performance and scalability, and yank out a system superior to anything in the market right now. I always thought it would be a Compellent or a 3PAR that would have done this. They both applied novel technologies, but within the context of traditional architectures.

OK, that was indeed progress. But an "infinite" storage system based on Fibre Channel could keep Fibre Channel in the game for a very long time. It is not uncommon for the purveyors of Fibre Channel storage to cry on my shoulders about how poor iSCSI performance really is, and how it is faking its way into the data center. To that I say, "bull." As I have indicated many times before, it wasn't the iSCSI protocol that got iSCSI moving, it was the way the protocol was implemented in a clustered fashion. The ease of use so commonly associated with iSCSI has nothing to do with the protocol. It has to do with the novel way in which it was implemented.

I believe the first major player that develops and ships a Fibre Channel equivalent of clustered block storage will run away with the prize. But, it will come at a price. The new storage will indeed impact the monolithic storage (EMC Symmetrix, IBM Shark, HDS Lightning) revenues for a while. But, as we all know, in technology business, we either eat our own children or competition eats them and us, in one gulp.

So who will be the first one to take this major step? The gauntlet is thrown!

About the author: Arun Taneja is the founder and consulting analyst for the Taneja Group. Taneja writes columns and answers questions about data management and related topics.


There are Comments. Add yours.

 
TIP: Want to include a code block in your comment? Use <pre> or <code> tags around the desired text. Ex: <code>insert code</code>

REGISTER or login:

Forgot Password?
By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy
Sort by: OldestNewest

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to: