Both And: Why the cloud and clients won’t work

I really hate the expression "both and". I was in a meeting once and asked, so what do you think is better for your company, Option A or Option B? And someone piped up and said, "both and" in a smug tone - I nearly jumped across the table. See, the question was about picking one or the other. There is rarely a "both and" answer, especially in technology, at least not in the real world of technology.

Let's say you are in a data center and someone says, "in your next data center, do you think you will continue using air cooling or move to water cooling?" and you reply "both and". Well, that's not going to work, because water-cooling is very expensive and needs to be installed as a priority method, a strategy. You can maybe add spot-cooling, and you can use some air cooling in the room, but as a long-term strategy, you need to stake your bets on one or the other.

I have this same issue with the cloud versus clients. As an industry, if we are going to build the cloud, we need to stake our bets that the cloud can handle most computing eventually. Otherwise, it is not really worth moving to a cloud model. Also, if we really believe in green computing, we will figure this out, because one mighty data center in the sky can handle issues like LEED certification and power distribution better than 1000 individual data centers in a state. Lastly, we're all just too bogged down with the client: installing them, supporting them, updating them.

I don't really agree with what Dan Lyons said today about Web 2.0, but I do agree with the sentiment that computing needs to solve more complex issues than finding good restaurants. Maybe you are against the cloud, but I see it as a way to bring utility computing into the mainstream and free up a lot of IT folks so they can do more important tasks than tweak a network switch for minor point upgrades. Maybe if we have the cloud, we can actually use computing resources to, I don't know, make a 3D map of the brain or cure brain cancer. Isn't pooling resources always better than everyone re-inventing their own wheels?

Mary-Jo Foley suggests that, at least with Microsoft and their software+services model, it's possible to do "both and". She says this strategy is not just a way to keep the enterprise from moving to the cloud and keep using Microsoft Excel at the same time. She notes:

It wasn't until today that... [with] offline software capabilities on phones and PCs ... (especially ones in emerging markets and with ‘occasional' connectivity) [using] services even without constant Internet access - that I felt that mixing the cloud and the client was more than just a rationalization strategy for Microsoft.

My thought: that's a great idea for Microsoft. It's brilliant to strategize that way. And Google, Cisco, sure - they can do "both and" and increase revenue on both fronts. But for the enterprise, clients are like air-cooling and the cloud is like water-cooling. You can possibly use both for a while, but you'll need to pick a strategy eventually.