Distributed computer networks control, scheduling and management are well-studied topics in scientific and engineering applications, yet the problem of finding optimal approaches that are not too complex to solve or implement on practical systems remain unsolved. Many computer networks are naturally clustered, and imbalance of load usually occurs among the clusters rather than inside the clusters. This project will develop clustered computer networks' control, scheduling and management policies for allocating computational resources across network clusters and for managing network operations based on simple network management protocol(SNMP). This project will first develop clustering model based on the data collected from monitoring Bradley network loads from real applications. Then the project will investigate and develop adaptive computer algorithms and heuristic schemes for developing the control policies. Finally, the project will develop network load monitoring tools based on X- window and Tcl/Tk for network manager interface. The control schemes developed are expected to be applicable to network load balancing and management in many other clustered distributed systems. The project is funded under the RUI program.
