What I Heard
I heard a few lectures about cloud computing this morning at ISOC-IL. Cloud computing is an evolutionary step up from grid computing. Essentially, your data and processes are replicated on many different locations online, and someone else manages (at least part of) your application hosting. You work on something in your office, and without any effort on your part, it's also available in your home when you get home.
One benefit is giving up your IT department (except to manage your PCs) and the costs and headaches this entails. A related benefit is that you pay for the capacity you need when you need it, without having to plan, running short on capacity, or paying for more than you need. Plus auto-backups.
Drawbacks are possible lags, service, security, compliance, and licensing issues. And all your legacy apps, of course.
One lecturer mentioned that, just like everyone stopped generating their own power 100 years ago and moved to the electricity grid, people will stop running their own computers and move into cloud computing.
What I Thought
At first glance, this sounds good. It is easy for a start-up to have access to high-powered servers without having to invest anything; that's good sense. I think cloud computing will grow in the near future.
But ... I've heard this story before, too.
Once we had very expensive computers, and everyone had to go to a cathedral-like server center to get any work done. Then we had cheaper computers and everyone owned their own computer. Then we had cheaper computers, and we all used dummy terminals that connected to a central computer. Then we had cheaper computers and we all owned a PC. Now we have cheaper computers and allow major companies to store all our data and services.
Cloud computing seems like a nice answer for problems we have today (IT depts are a nightmare), but what about tomorrow's needs? What happens when computers cost $0.30 and bandwidth is effectively infinite and free?
With cheap computers and the right software, we can own grids of disks and processors where today we own single servers. When one dies, our replicated data will live on in dozens or hundreds of our other cheap disks. We'll just pull it out, toss it, and stick in another. Same with processors.
I see our IT departments as expensive today only because our computers still justify the cost of this expense. When computers are as cheap as light bulbs, assuming we can run open source cloud computing software, IT will not really be a problem. What will become of these cloud computing companies when we no longer need them?
Maybe I have a pie in the sky attitude about this, but that's what I envision. Cloud computing looks to me to be one baby step on the road to my vision of Web 3.0.