On the Implementation of Imago System
Xining Li
2006-01-01
Abstract:Mobile Agents are mainly intended to be used for network computing -applications distributed over large-scale computer networks. An intelligent mobile agent is a selfcontained process, dispatched by its principal, roaming the internet to access data and services, and carrying out its assigned decision-making and problem-solving tasks remotely. In this paper, the author will present the design and implementation of the IMAGO (Intelligent Mobile Agents Gliding On-line) system. The goal of the project is to build a logic-based framework in the design space of intelligent mobile agent systems. To achieve this, we need to cope with design issues, such as explicit concurrency, code autonomy, security, communication/synchronization, service discovery, computation mobility, as well as implementation issues, such as multithreading, garbage collection, code migration, communication mechanism and database access. The unique feature of IMAGO system is that it deploys intelligent mobile messengers for inter-agent communication. Messengers are anonymous, thin agents dedicated to deliver messages. Like other agents, messengers can move, clone, and make decisions for their assigned task: track down the receiving agents and reliably deliver messages in a dynamic, changing environment.