Day 1. Nothing clever and nothing new
Here’s what’s clever about cloud computing: it’s nothing special and nothing new again and again.
I’ve run two servers for years.
Each client get two servers – or 4 servers – or 7 servers. Some are databases, others web, others do other stuff. But there’s always the same number of servers.
The there’s cloud servers.
The important thing – they tell me – is to make each server disposable. No server must matter. I must be able to stick an axe in it, and you’ll be able to survive.
And this gives you two things: you can go up, and you can go down.
But wait, none of that makes sense. So let’s start again.
You’re website runs on some server, but it doesn’t really matter which server. When the page is built for one of your esteemed visitors, some server pulls the data out of the database, builds up a load of HTML and throws it back at them.
It cloud architectures, it’s important that you can create any number of servers to do this.
The key is to work from a very plain server setup. Take standard ubuntu installation, configure it, add your code and make it like. If configuration takes hours and requires a human to get involved, albeit a nerdy human, then you aren’t really running on the cloud.
The way we’re approaching this now is to create new instances that are very plain, send a script to the new instance that sets up the internal DNS, installs all the required software, configures it and grabs our code.
So when our servers run slowly all I need to do it… all that stuff, but again. Nothing special and nothing new. That’s what’s clever about cloud computing: it’s nothing special and nothing new again and again.
