A step backwards for a moment
orbist 060000HPM5 Comments (3) Visits (3649)
A work colleague asked if is RSS feed for my blog was broken the other day. Things have been sure crazy for the last few weeks - thats always my excuse! I did have this post in the sidelines, and wanted the mud slinging to stop before I set the record straight. Well, you can probably guess who's blog in particular I am referring to as soon as I mention SSDs...
I kind of knew it would be coming, it was bound to. IBM hadn't made any public statements about our research and development efforts around SSD devices, maybe I need to clarify that with Enterprise SSDs. Obviously we had our Quicksilver technology demo from last year, but other than that we'd not really said anything about where and how we'd be using them. Certainly February's IBM Pulse based announcements created some heated debate out here in the blogs, some "science experiment" numbers from EMC (Barry Burke even had the audacity to publish number's he'd got from 'testing' a DS8000 - pitty they won't do the same for their own products - eh?)
Anyway I wannted to get a few things straight, so here goes :
What we need to make best use of all the IOPs and the response time is a flexible, scalable, high performance architecture, which is exactly what SVC is - and exactly the point we were making with the Quicksilver technology demo. Now I agree, that was a demo, but it doesn't change the message. Today we can all add SSDs into our existing products, the Enterprise boxes being the first to benefit. These are the environments that need guaranteed response time, and today are often short stroked. This is also why SSD update will be slow. Its only those people doing those kind of response time critical operations that will benefit, and can afford the current pricing. I'm not going to comment on the list price of the DS8000 SSD options, if you are interested then speak to your account manager, or sales team and they can discuss.
Since our Quicksilver demo, my inbox has been running hot with requests for more information, and those customers we've spoken to (who come from all industry sectors) have a wide variety of applications that are looking for not only response time improvements, but also huge IOPs demands. Rest assured when we do release SVC and SSD support all of these needs will be catered for.
Lets close by coming back to the title of my post... a look back at 2007, where if you believe everything BarryB says, they were alone in their work on SSDs...
Edited bullets bellow to clarify my statements, correct where necessary based on BarryB's comment below
So for all the noise... being first isn't everything, yes they probably helped to kick start the discussions and accelerate early adoption... but I've read many reports that says 2009 could be slow for Enterprise SSD adoption. Its going to need more than just the raw device support, and thats where we come back to my fellow IBM bloggers 'supermarket' analogy. IBM can be a one stop shop for all your IT needs, it also means research and development from software (applications, operating systems, management), hardware and services can all discuss, design and produce an entire Information Infrastructure that is not only Dynamic, but can also make the very best of a new and exciting technology such as SSD.
PS. The other debate that ensued, and it seems certain bloggers just can't agree, but for once I do agree with BarryB, there is a need to distinguish between the "laptop" SSD families as I call them, and the true Enterprise class SSDs. The devil is in the detail, and its this detail that makes only certain SSD devices applicable in an 24x7 storage environment.