A method and apparatus for providing scalability and fault tolerance to a distributed network. According to one embodiment of the invention, a network management system comprises a set of one or more element management servers. One of the plurality of element management servers is designated as a master server. The master server determines which of the plurality of element management servers are to manage each of a set of one or more network elements. In one embodiment, peered and federated services are used within the network management system to provide scalability and fault tolerance.