Efficient dynamic itinerary and memory allocation for mobile agents

Abstract

Recently, the distributed and mobile agent concept has become a new computing paradigm in the Internet distributed computing. Itinerary planning is one of the important requirements for mobile agents in completing the given task efficiently. The static itinerary planning may not be the best approach in real network environments as the status of network components are subject to change dynamically. Dynamic itinerary planning makes the agent to change the itinerary dynamically according to the current state of network. For an information retrieval application, the main issue is to find an optimal itinerary for the mobile agent in order to achieve the efficient and effective data collection. Further, for a host to accommodate a remote agent, it must be ensured that the work load of the host is manageable. This paper provides an environment that selects the next optimal host for the mobile agent to visit by requesting the available directory services on the fly. For a host, to accommodate the incoming agents and to maintain work load, this paper provides a mechanism that either destroys the agents for which the task is completed or transferred to secondary memory if the mobile agent is waiting for an event to occur. The presented work is tested in a web environment and the results obtained are more encouraging in claiming that this work provides an efficient dynamic itinerary plan for the migration and sufficient space at the host for the execution.