Introduction

Abstract

An appealing feature of next-generation information agents which is attracting increasing interest is mobility. Roughly speaking, mobile agents can be transported to different sites in the Internet and provide mobile data access and processing. Contrary to popular belief, mobile agents do not really move themselves, but depend on a given mobile agent system for this purpose. A mobile agent can be seen as a continuously executing program only interrupted during the transport between several servers. Such an agent is almost written in an interpreted machine-independent language such as Java, so it can run in heterogeneous environments. It is assumed that an appropriate computation environment is accessible on any server the agent might visit.