An intelligent load distribution system for CORBA-compliant distributed environments

Dynamic invocation of services in distributed O-O execution environments requires effective techniques to balance the computational load among available servers. In this paper a fuzzy classification unit is proposed as the basis of the design of an intelligent load balancing system for distributed O-O servers in the framework of the CORBA reference model. The proposed system is based on CORBA's standard Naming Service and relies on a fuzzy rule base to take into account the different performance indexes of the machines hosting the servers and the fluctuations in the quality of the network links among them.