In separate posts, Jeff Kaplan and Phil Wainewright make a compelling case that 2008 will be the year when web apps and other utility computing services move into the business mainstream. They offer a slew of reasons – 18, to be exact – to back up their predictions. I’d add just one more: A lot of suppliers have been investing a lot of capital into constructing utility data centers over the past couple of years, and a good deal of that new capacity has either come online recently or will come online during the course of this year (including Microsoft’s huge Chicago plant, which is scheduled to open in April). It’s important to remember that we’re still in the build-out phase of the utility infrastructure and that supply-side capacity will be as essential as user demand in pushing the model forward. Capacity, as well as demand, should expand significantly over the next 12 months.
It has less to do with the physical loci of data centers and more to do with advanced programming language bindings and libraries that make it easier to use these systems.
Existing data center hardware can be converted, and superserver clusters are becoming more common. And, then there is the virtualization revolution that has not fully become realized.
In summary: It’s not purely a matter of physical provisioning, but depends more on tools integration.
One obstacle that a lot of SaaS plays have yet to tackle is how to make money. Cheap hosting, wide pipes, and inexpensive development can only go so far. At some point advertising revenue just doesn’t scale.
Nick,
The break out of microprocessors for embedded uses led to the PC. That is, latent in the new economy of CPU power was a platform that was, in and of itself, a mass market product.
Cloud computing is another crank of that dial — a new economics for CPU power, with new ways of packaging CPU power. As with microprocessors, the first applications are for business use.
Will there be a breakthrough in “personal cloud computing”? Will people one day provision personal servers the way, today, they provision email accounts?
(My view.)
-t
It would be interesting to revisit this prediction in a year. To me this looks more like a milder version of 1999-2000, when all the telcos were looking at the traffic growth charts, and frantically laying down fiber. What we got out of that? A nice stock bust, couple years of doom and gloom, and loads of dark fiber.
I think utility computing will eventually happen (just like all that dark fiber is being gradually lit up now), but we are not quite there yet.
Twka90,
We got a lot of *highly contended for* dark fiber. The bets are still on the table, is all that’s going on.
-t
Tom, I’m sorry, are you saying that dark fiber is not being lit even now?
Twka90,
Uh… I dunno. I assume so. Did I miss something?
I took you to be trying to draw an analogy between commodity computing stuff being built out today and cable lay-downs in the past. Cable remains dark, I take you to say, which proves you can’t bet much on the basis of merely the evidence that build-out is occuring. Is that right?
I’m saying that there mere fact that a lot of cable remains dark is no proof that it was poor investment. You’re looking at a different horizon than the investors, is all.
Did I just entirely miss your point and respond non-responsively? (If so, sorry.)
-t
Tom, no, you are absolutely correct — “build it and they will come” does not necessarily work. The fact that Google/MS/etc. can offer 10GB email accounts for free does not mean that SaaS is going to take over… It would be interesting to watch though.
Personal cloud computing! What a great idea! Hope it happens or SaaS will always be a B2B or strictly a programmers paradise. Big but still on the fringes – backstage operations! Marketing push will drive costs but still leave masses out in cold and wondering what’s in it for them. Slow adoption.
There has also been talk that SaaS (On-Demand) is ready for a fall. One of the ways to watch the SaaS market is to watch the stock performance. At The Data Doghouse blog I’ve introduced the On-Demand (or SaaS) stock index as a way to keep an eye on it.