I’ve been thinking about this a lot lately, and figured I’d toss out some alarming concepts to stir up a discussion.  We’re researching the topic and will publish a more complete paper on it shortly.

The premise: it is only a matter of time before disk (as we know it currently) will stop being a significant part of the data center.  Instead, solid state/flash type technologies will replace it.

What time?  I think within 5-8 years.

Why? Because the servers are solid state.  The networks are solid state.  The only thing that ISN’T solid state is storage.  The only REASON that continues is because of the cost.

Previously we wrote a paper on the economics of change–where we (rightly, in my opinion) stated that it takes a step function in cost reduction to move a market to a new way of doing things.

Few argue (although there are some, which I always am thankful for) that mechanical disks are BETTER than solid state technology.  People did argue that vacuum tubes were better than solid state stuff too.  How did that work out for you?  People run mechanical disks even though they are slower, inherently less reliable (more on this argument in the paper), and generally a nightmare to manage (due to the sheer number of them) simply because the cost of the alternative is just too high.

Thus, the contention is that IF the cost were on par or at least close, people would choose solid state devices to store their stuff on versus disk drives–at least for primary capacity.  (Eventually EVERYTHING will be solid state, but just as tape maintains a role, disk will most likely maintain a role for a long time–perhaps for archive systems, etc.–although if cost truly becomes a non-issue, then one can draw the conclusion that all we need is solid state and tape.).

Yes, there is a big IF–but since that IF is economic–I can assure you that issue will be solved.  Then it just becomes about WHEN.

So, if it is inevitable that the end game = solid state storage in the data center (everywhere), and that end point occurs at X, what are the likely disruptive opportunities that will avail themselves to our IT industry along the way?

Dick Egan of EMC fame told me and others, in 1986ish, that he felt that by 1990-1992 we would no longer use spinning disks. He was off by a few decades, but he was right.  Moore’s law doesn’t apply to disk drives.  It does continue to apply to solid state technologies (like processors, memory, etc.).  Couple the natural market forces with new acceleration technologies such as dedupe/compression and one can see how in relatively short order, the economic argument will be flushed down the toilet.

I’ll focus the next bit on what the likely progression will be in getting from point A to point Z shortly.

IRRELEVANT SIDE NOTES:

• The NHL playoffs were one of the best sporting events I’ve seen in a long time.  Of course I dig the fact that the Bruins won it after 87 years, but it was awesome to watch it all regardless.  I shall be visiting the cup this Sunday in Edgartown, MA.  Why it is going to be there, I have no idea.
• I’m very excited to go see the reunion show of one of Boston’s legendary local bands, the Del Fuegos tonight, where I shall regale in tales of my much misguided youth.  Back in the day I was in a great punky type band that played with them a lot.  I wanted to be a rock star, but thankfully for you I became a geek/business voyeur instead.  Chicks don’t dig me anymore, but at least I can afford stupid things now.
• I smell another big recession coming on in Tech.  I hope I’m wrong, but the hair is standing up on the back of my neck.  China is slowing spending, Europe is a mess, and the U.S. government has to stop spending like drunken sailors while using credit cards to pay of its other credit cards.  It all feels like we could be heading for a problem.

Cheers

Related posts:

1. Big Problems on the Horizon: Disk Shortages
2. Was Dick Egan Right After All?
3. Holiday Thoughts – and Disk Drives
4. Is it Time to Reclassify Storage?
5. Scarcity Continued….