Plain Scaling

Getting up and running

The intention here is to talk about things that most people aren’t going to think about. How deadlock prevention code in your database may start to bite you at certain levels of concurrency. How your HTTP stream breaks down into packets, and then frames on the wire. Data decomposition for parallelism. Algorithm decomposition for driving you crazy. There’s a lot of good commentary on this already, both in books and on the web; it may turn out that there’s nothing interesting to say here at all.

Of course, I might get distracted. That’s happened in the past - but currently I’m actually being distracted by something scaling-related: Tim Bray’s Wide Finder 2. I have a bunch of posts about that on my personal blog. Really they should be here; really really they should be syndicated to here from there, or vice versa. Atom has support for that, but I’d probably have to go out and write some software somewhere to get it to work…