For some time now I had been wondering how applications like Google search and Facebook with millions of users manage to serve requests almost at a fraction of a second.
I am not aware how it is done in Google search but today I was lucky enough to read this pretty interesting article on how this is done in Facebook.
According to this article the Engineering crew at FB has managed to scale memcached, a high-performance, distributed memory object caching system to handle 200,000 UDP requests per second.
That is really awesome !!!.