ORIGINALLY POSTED 15th August 2009

13,321 views on developerworks

I, like a lot of you that take the time out of your lives to read blog ramblings, you probably have your top 3 or 4 authors that you will read no matter what they say – maybe the next 10 that you read just the title, and based on the grab line, may or may not read, then the others that may simply read when you have nothing better to do.

I was recommended a book by fellow IBM Blogger Tony Pearson when I started spilling my thoughts, and it was well worth the read on a state side trip. “naked conversations” by Robert Scoble and Shel Israel. One of the useful tips was to write the title first. At the point we bloggers decide there is a post in our thoughts, generally you are het up and have something to say, therefore thinking about a quick and concise description for what you are about to say is the best place to start. One you start off on the meandering of blogging, your thoughts get tainted and the purity of the initial thinking is gone.

One of the bloggers that fit into my ‘tier 2’ category is our HDS friend Hu – now that is not meant in any way to be a punt at Hu, but more of his writings – I too get dragged into the Marketing side of our business, but I try to keep my ramblings my own. I don’t for example look for ghost writers.

Hu’s recent post titled ‘the changing storage pyramid’ really did make me question – exactly how is this ‘changing’ ? Isn’t this the same as what we already have ? Maybe its more thinking along the links of – ‘Hitachi Maths’ – many would like to claim that they coined the phrase (I will attribute myself to claiming the phrase ‘doing a netapp’ wrt blog polls) but I do have to attribute my namesake BarryB as the maths (or should that be “math”) phrase finder.

I have to completely disagree with what Hu thinks. I’ve been playing with enterprise SSD devices as long as most. Having the advantage of getting interested in the technology at the right time, and having a boss who could see the value of looking into this upcoming technology. I see the SSD, HDD and tape market as much more of an egg timer shape than a pyramid. Yes so today, a few key players own the market and are charging for their IP, but is the underlying physical media that different from what we have as a commodity today… i.e. disk, or more realistically storage capacity.

On the one hand we have spinning electrons on some rust inside an aluminium case and on the other we have some spinning electrons on some silicon inside an aluminium case. Either way, rust and sand are pretty common elements. The hardware itself isn’t expensive. Therefore I have to think that as we get some serious competition (they will all use Samsung silicon parts…) but the vendor competition is needed to push the price down.

So here is my picture – and of course the pinch point will be the traditional ‘enterprise drive’ i.e. the 15K RPM Fibre Channel disks.

Please excuse the pun, but I do believe that SSD, or flash, will be a ‘flash in the pan’. Its not going to be another 50 year old technology like HDD was, there will be much more reliable technologies, longer term reliable that is, phase change memory, racetrack, whatever they may be. But for now, flash is real, and is here. Some clever folks have worked out how to make it last at least 5 years in an enterprise situation, and as more and more vendors make SSD drives, so the price will come down. Its already hitting the real commodity consumer market. I’m not saying OCZ’s SSD drives are on the same par as STEC, but it won’t be as long as its taken us to get from 1MB HDDs to 1TB. (Some 45 years).

The pinch point really is the high performance HDD market. Hu has not seen what I obviously have. The days of short stroking many HDDs are gone. The power bill along means that you can now justify that bill for one or two SSDs. If only you got to squeeze every last IOP that they *could* provide… more to come

Comments from post on developerworks: