Location independence for mobile agents

In recent years there has been an explosion of interest in wide-area distributed applications, executing on intranets or on the global internet [CHK97]. A key concept for structuring such applications is that of mobile agents, i.e. units of executing code that can migrate between sites. Mobile agents require novel programming language support and raise new problems, of interaction between agents, of robustness under network failure and recon guration, of binding to resources, and of security. In this paper we study the rst of these, considering the design, semantic de nition and implementation of communication primitives by which mobile agents can interact.