Agent oriented modeling for proxy server

Abstract

Intelligent proxy server is one of the new technologies for firewall system. This type of proxy can do a great deal more than simply relay requests. They use agent to implement caching concept in order to get better transmission between client and server like CERN HTTP proxy and Microsoft Proxy Server. The agent was meant for search and information retrieval which implements caching concept and handle all request for remote documents. The proxy server itself is not an agent; developers developed the agent separately and incorporate the agent into the existing firewall. Compared to Proxy Agent in this research, the proxy server was developed as an agent and was meant for fixing the configuration problems. It follows from SIGAL's architecture, which based on clientlserver. The improvement towards the architecture has been made to fulfill the TIS FWTK proxy server. The type of agent used in this architecture has been changed into intelligent agent and to implement this, the CBR technique has been incorporate into the development process. Thus the agent has been programmed to have adaptation and learning features. To develop such multiagent system, a structured method must be followed and therefore MaSE has been chosen to model the Proxy Agent framework. This research has provided with a step-by-step modeling process for Proxy Agent and has been implemented and tested using AgentTool, which comes together with MaSE. The Proxy Agent framework has also been tested to be used in real proxy server environment and the result shows that the Proxy Agent has better performance in performing the configuration (network configuration) compared to manual configuration which always cause misconfiguration.