large scale distributed website usually involves with the distributed databases ,files systems,and cache systems.the infrastructure is built based on features like the huge traffics and parallel operation etc . the large scale website need the support of the infrastructure which has high availability , manage ,highly performance, Scalability, maintenance
the website can be subdivided into sub systems,like images system, post system, new systems .with the load balance methods dispatch these requests to the different sub systems
in order to let different sub systems share the same server ,avoid the single point of failure,unified the configuration ,centralized the management. we should unified all the servers ,one server can deploy many sub systems. when traffics become bigger and bigger, we can add the nodes of server to improver the performance ,that is cluster
http://write.blog.csdn.net/postedit