Three riffs on the cloud

I wuz hacked! Call the police!

Oh, wait. There aren’t any.

It’s funny, in a not-at-all-amusing sort of way, how we’re able to maintain mental models of the world we know aren’t true until someone rubs our noses in the disconnect. The hack job on the KJR archives is a perfect example.

My mental model of the world: Someone commits a crime. The victim calls the police, who investigate and, more often than not, catch and prosecute the perpetrator.

How it really works: If we want enough police for my mental model to work, we’d better open our checkbooks wide to our local taxing authorities, because whether the subject is stolen bicycles or hacked blogs, the local gendarmes aren’t going to do anything more than take your name and express their sympathy unless we hire a lot more of them.

KJR is hosted by a cloud vendor. It provides a nice little toolkit for building websites and also the tools needed to install a WordPress instance, which means it qualifies as a PaaS vendor (platform as a service).

And as I always knew, only this time it was personal, putting your infrastructure in the cloud doesn’t make it someone else’s problem. I had to figure out how to clean up the mess, how to secure my archives better, and how to get Google to take down its warnings.

* * *

Whether you’re a cloud evangelist, skeptic, or somewhere in between, read “Why PaaS? Dev, test, staging, no waiting,” (Andrew Oliver, InfoWorld, 5/30/2013).

The business case for the cloud has been foggy since the technology’s inception. It’s rested on “advantages” like cloud spending coming out of the OpEx rather than CapEx budget. This never made any sense, but especially it makes no sense at a time when, in the aggregate, the business community is sitting on large cash reserves because most companies don’t need more employees and can’t figure out any better place to invest them than stock buy-backs or the bond market.

Nor is cloud computing necessarily cheaper than its data-center equivalents. It’s more flexible, but not cheaper: Cloud computing lets you add and shed capacity as you need it.

What Oliver adds to the discussion is a special dimension of capacity management, namely that for most businesses, the capacity needed for development and (especially) test environments is occasional rather than fixed.

Also: In traditional computing environments, managing development and test environments is, shall we say, a non-trivial task.

But with cloud providers you can spin up copies of your production environment quickly, and relatively affordably because you only pay for the capacity when you need it.

And, done right, you don’t have to migrate test to production – just switch a DNS entry and test becomes production. Then you spin down your old production environment and go on your merry way.

What’s particularly likable about this aspect of the cloud business case is that it fits the KJR model of the world: Most of the time, what matters aren’t lofty, vague, strategic visions. What makes one business successful while another fails depends, more often than not, on the basic blocking and tackling. And in IT there isn’t a lot of blocking and tackling that’s more basic than change control.

* * *

Back when the earth was young and I had more of a use for barbers, there was this innovative product called the personal computer. It changed things. A lot.

But something it didn’t change at all was the business need for mainframe computers. What PCs did for businesses was to take on tasks mainframes had a hard time scaling down for, like electronic spreadsheets and word processing.

But PCs also opened the door for distributed computing, which did reduce the business need for mainframe computers. The progression was PCs, to PC-oriented LANs, to LANs connecting PCs to servers and minicomputers, to “the network is the computer.”

I’m starting to wonder whether we have a similar dynamic going on right now with mobile computing and the cloud.

See, mobile computing, properly understood, is more than “we need to support smartphones and tablets.” It’s a matter of making a company’s computing resources available to everyone who needs them and should have access to them, wherever they are and on whatever device they’re using, in a presentation well-suited to that device.

So I wonder whether mobile computing is opening the door for the cloud in much the way the PC opened the door for distributed systems, because one of the cloud’s virtues is that it can deliver your applications wherever they’re needed.

Yes, an analogy may not be the same thing as being the same thing (creds to the Economist), but exploring analogies can provide useful insights.

Think this is one of them?