icis99 abstract

An agent-based computing model has been proposed as a new computing model with many useful potentials for distributed computing over the Internet. Agents have also been used extensively for constructing intelligent applications. While a lot of work has been done in designing intelligent agent, efforts in providing system-level support to deal with component failures and security threats have been lacking. This paper describes the design and implementation of an agent-based computing system called DaAgent. DaAgent has been specifically designed for experimenting with various fault-tolerance and security mechanisms in an agent-based computing system. A prototype implementation in C indicates that system-level support for fault tolerance and security can be provided at a relatively small cost.