SNW & NAB: IOPS vs bandwidth

by Robin Harris on Wednesday, 23 April, 2008

NAB frame by frame
SNW and NAB did not overlap this year, so I spent 3 days each at both. The 2 events are very different: storage is the topic at one and merely central to what everyone is doing at the other. I enjoy both.

Rather than tackle NAB in one piece I’m writing a series of short takes on a number of companies.

SNW is the past. NAB is the future.
Storage is in the midst of a massive transition from an IOPS focus to a bandwidth focus. Like computing’s shift from batch to interactive in the ’60s and ’70s this transition is about bringing the technology closer to how people live. Not consumerizing, humanizing.

Life is a sequential access workload. Our eyes, ears and our pattern-hungry brains crave bandwidth. New display technologies push patterns at us at rates that used to require roller-coasters.

Batch isn’t going away – Google probably runs more batch jobs than most F100 firms – and neither is transaction processing. But the investment goes to the growth areas and bandwidth intensive storage is a growth area.

HD 3D: the Next Big Thing?
3D is getting good and will be the next step in home theater. Whether it is good enough to break through in theaters is another issue. But the net is that high quality 3D doubles data rates.

NAB gets this. They also get that to be useful, storage has to be integrated with the application, whether that app is production, editing, distribution or presentation. A new wide world is opening up to people who know storage and can learn an application. Much easier than the reverse, to be sure.

And, of course, they have a very reliable market to pay for all the innovation: entertainment. Cool.

Comments welcome, of course. First up: Isilon.

{ 1 comment… read it below or add one }

Charles Soto April 24, 2008 at 6:39 am

This is a Good Thing(tm). Storage as a “market segment” has always been the wrong focus. Storage is little more than a place for servers (and to a lesser extent, workstations) to keep their stuff. I’ve been very happy with the “storage leaders” embracing software differentiators, as its data management tools that bring real value (ASICs and spiffy backplanes are run of the mill). Unfortunately, storage leaders got used to storage premiums and sell that software at a premium.

Meanwhile, server teams have already been through the commoditization phase and see software as a means to drive sales of their servers. Once again, we’re seeing general purpose servers taking over many of the duties we once required expensive software on expensive arrays to perform, and they’re mostly doing it with open standards. As a result, commodity storage built around open standards (iSCSI, pNFS, CIFS, etc.) are going to explode. This will allow us to put money back into the real value generating solutions – things that improve workflows in our respective industries (I manage IT for a college of communication, so we have a lot of broadcast/film).

I look forward to when “storage” isn’t really considered enough of a specialization that it requires dedicated blogs to follow. Not that your blog isn’t appreciated 🙂

Leave a Comment

{ 1 trackback }

Previous post:

Next post: