Mobile Agent Systems Execution Code on a Mobile Agent. VS. Execution Code on a Stationary Agent.

الملخص

While many researchers investigated the merits of the mobile agent paradigm over the conventional methods like Client/Server and Code on demand, very few provided an insight within the software agents community itself. Such an insight is provided in this paper. Within the context of the software agents two categories can be identified: stationary agents that execute only on the system where they begin execution (Home) and mobile Agents that are not bound to the system where they begin execution. A mobile agent is firstly residing on a home machine, and it is dispatched to a remote host for execution. Like any other computer program, a software agent needs code to execute. This code could be installed on the stationary agent at the server side or it could be pre imbued to the mobile agent from the client side. The placement of the execution code is the main objective for this paper .Using Aglet Software Development kit (ASDK) two states are studied, first where the execution code pre-imbued into the mobile agent and second where the execution code is preinstalled on the remote server and manipulated by the stationary agent on that server. In terms of performance, single mobile agent is compared to a stationary agent, and multiple mobile agents are compared to a stationary agent.