A system and method for server load balancing that includes server affinity. The system can be incorporated into application servers to support load balancing algorithms for RMI objects that provide server affinity. Server affinity uses smart load balancing for external client connections in such a way that it uses preexisting connection and reduces overhead: the client considers its existing connections to the application server instances when choosing the server instance on which to access a service. If a service is configured for server affinity, the client-side stub attempts to choose a server instance to which it is already connected, and continues to use the same server instance for method calls. All stubs on that client will attempt to use that server instance. If the server instance becomes unavailable, the stubs fail over, if possible, to a server instance to which the client is already connected.