HAProxy is easy to set up and is probably what he wants, if he plans to scale to multiple machines. But on those machines I'd recommend cluster.js. (http://learnboost.github.com/cluster) It will spawn n child workers, proxy requests to them, and restart them if they die.
Comments
HAProxy is easy to set up and is probably what he wants, if he plans to scale to multiple machines. But on those machines I'd recommend cluster.js. (http://learnboost.github.com/cluster) It will spawn n child workers, proxy requests to them, and restart them if they die.
It's worth mentioning that clustering will also be built into Node.js 0.6.0 which is being released in the coming week according to the official blog.
Don't know if it's as extendable and pretty as TJ's cluster though.