Every once in a while you hear something going on in the political spectrum that strikes you as meaningful and useful, that you hope against hope they will manage to get past the details and partisanship and move forward on. Right about the time of the writing of this blog, in the United States, we’re hearing a lot of these things because it’s an election cycle. Problem is that with 300 million people it is rare that any one idea is agreed upon by everyone, and politicians cover the spectrum, so often what sounds like a good idea is tough to get pushed through, and election years tend to generate more ideas and less action than other years.

Every once in a while you hear rumblings in the industry that are in line with what customers want, and you hope against hope that they’ll overcome natural vendor bias and differences, marketing initiatives and the chance to win market share to actually get it done. During tough economic times you hear more of this stuff as companies struggle to maintain their market share – or grow it – in the face of shrinking budgets. The problem is that with hundreds of vendors, all competing while trying to develop interoperable products – as is the case with storage or networking – it’s tough to get everyone to agree on the right direction to take to improve the products on the market. Usually, one vendor is successful at some new wing-ding, and the others follow on.

I honestly didn’t believe that the storage industry would move toward a converged network infrastructure. Seriously. There were companies out there head-nodding to it, and some making solid product in that direction, but the underlying current of “not for tier one” or “not ready for prime time” was always there with the big guns. At least I heard that, I’m guessing you did too.

But then I happened upon a blog post by Chuck Hollis of EMC last night – I Want My pNFS and I was thrilled to hear him say that the market was moving in the direction of converged networking. More thrilled to hear they’re getting together for a plug-fest. This stuff takes time, you don’t make a ton of vendors play nice together overnight when the devil is in the (packet) details… But they’re moving in that direction.

I’ve said before that what the storage industry needed was one protocol and an end to a stream of them that are good for some vendors, bad for others, and end up with you and I having to deal with at least two protocols in our storage environments. I also said in that same post “sorry storage vendors, you shouldn’t be the people to do this…” because of the inherent conflict of interest in setting up a unified storage infrastructure that gives customers adaptability in moving from vendor to vendor, and the long/storied history of storage providers and vendor lock-in.

But pNFS and the post by Mr. Hollis may well have changed my mind on that last bit. If the vendors are ready, willing, and able to get a compatible converged network up and running, and it utilizes a single protocol to access data… Well, that’s just peachy with me. I admit surprise, but it is probably the best solution if they can pull it off.

Now sure, you might be thinking “You have ARX, any file-level storage protocol is going to be fine with you – you can sell more product”. That’s true, but notice that I didn’t say if I cared that it was file-level. I think file-level makes more sense because products like ARX can optimize and distribute your disk usage via policy without resorting to making your data unrecoverable, but I’d be nearly as thrilled with block-level, as long as the storage industry had a direction other than increasing fragmentation of protocols. In short, my interest is much bigger than whether F5 sells a few more ARX boxes, it is in making the enterprise storage environment less complex through merging of usage to a single protocol – or at least a single physical layer. Ten Gig is definitely fast enough to host 99.9% of all storage needs, so it was a question of the vendors giving you what you need.

And hopefully this effort means they’re going to. No guarantees of course, vendors can cling tight to their current implementations and offer this as an “add-on” that’s not well supported – like many did with iSCSI –  but let’s hope it transforms the market rather than adds to the clutter.

By way of disclaimer, we’re a business partner with EMC, but that had zero impact on this blog, they just happened to be the vendor I saw talking about it.

“Transforms the market…” Hasbro’s Optimus

Related Articles and Blogs

• Given Enough Standards, Define Anarchy
• Pursuit of Intercloud is Practical not Premature
• HTTP Now Serving … Everything
• Store Storing Stored? Or Blocked?
• F5 Friday: THE Database Gets Some Love
• Load Balancers for Developers – ADCs Wan Optimization Functionality
• Cloud, Standards, and Pants
• Don MacVittie - BIG-IP
• DevCentral Interviews | Audio - F5