Supporting Disconnected Computing in Mobile Agent Systems

Mobile agents are typically referred as a very adequate paradigm for using in mobile and disconnected computing environments. They are autonomous, have a strong sense of location and are able to operate independently of a central server. Nevertheless, although this is true, for developing a mobile agent system that is capable of operating in a disconnected computing environment, much more is needed: persistence stores where the agents can be safely stored; agent reactivation mechanisms, which allow agents to migrate when a network connection is detected; policies and notification mechanisms for dealing with changing locations; among other things. In this paper, we present the mechanisms developed for the M&M mobile agent framework in order to address mobile and disconnected computing environments. These mechanisms can be readily adapted to any other platform that needs to deal with disconnected computing. Keyworks: Distributed agents, mobile agents, mobile computing, disconnected computing