It's that time of year again. MacWorld 2008 is just about to start
and already all of the liveblogging sites are clogged. Some are
completely down due to demand. Others force you do hit reload, which
fails 50% of the time. Why doesn't a good liveblogging client/server
Here's what you'd need and other random notes:
- A simple client library that does a poll to the server asking for everything since "last update" and updates the page appropriately (incremental). A little icon saying when next poll is, that a network connection is happing would be swell. If the "event" is over, then it knows to stop sending the request.
- The server is more interesting. First you need a way of publishing (writing) data and no matter how busy the server is, the publisher must be able to get in. This probably means two servers to be absolutely sure, but who knows.
- The ability to add more front end servers.
- A ridiculously fast HTTP server. As a goal a single server should do about 10k connections per SECOND. If you poll once a minute this hopefully will be able to serve 50k people per box.
- Event-driven http is probably the way to go since you are going to be I/O bound, not CPU bound, but you could start with anything server for kicks and swap out the server as you go.
- You could use fixed-size message block internally, say 1-4k, to make memory or file lookups easy. (e.g. append to the file, then the clients can stat the file and read from the end and go backwards to get the updates). Or something
- A very fast time out.. any client taking more than a second, or even less, should get the boot!
- You get the idea
It's a good project to work since you'll be able to work with all the
fun stuff you normally don't get to do fool around with. It's not much
of a business but who knows. Start with an open source and then do
consulting and then hosting. If anyone is hell bent on doing this,
let me know!
Note: CoverItLive appears to do this, but
uhhh, not so well, since it's front door says it's out of capacity