Security Issues in Agent Based Computing

With the increase in popularity of computer networks, there has been a shift proposed in distributed system programming from the remote procedure call (RPC) to the remote programming (RP) paradigm, to decrease network traffic and improve performance. To accomplish this, several experimental systems use agents (i. e., mobile processes) to implement the RP paradigm. The use of agents has several advantages and a few disadvantages, including added security issues. In this paper, we survey several experimental agent systems under development and describe their current security mechanisms. We then develop a general agent model and discuss general security issues in that model. Finally, we propose new solutions that address some of these security issues. Resumo Com a crescente popularidade das redes de computadores, tem sido observada a tendência de substituir o tradicional paradigma de chamada a procedimentos remotos (RPC) pelo de programação remota (RP). O principal objetivo é diminuir o ..