The problem of monoculture

I wrote a piece a while back about how centralized computing makes a cloud a big target. I didn’t want to get into the biological origins of this stuff, but one commenter was right: Monoculture is a precursor to extinction.

In university (which seems a long, long time ago) I wrote my thesis on evolutionary theory and product life cycles. Admittedly, not a screamingly fun topic, but it did give me a chance to read up on the Burgess Shale and other such things.

Now comes word that Amazon’s EC2, by virtue of the independence it affords hosters, is being used by bad guys for nefarious misdeeds (thanks to Rachel Chalmers of The 451 for pointing it out.) This provides an additional risk: Many of the Internet’s defense mechanisms involve black-holing specific hosters when the sites they’re operating do bad things.

Of course, when you’re hosting many applications, having one of them get blacklisted can be a nuisance for all the others. What’s interesting is the back-pressure we’re seeing arise against the popularity of cloud computing: At Structure, we debated the fear of lock-in; Stacey has a great piece on enterprise obstacles to adoption; and here, we’re seeing the downside of on-demand, easy-access platforms.

In other words, the bigger they are, the harder they fall. And that doesn’t just apply to dinosaurs.