Research Of Security Mechanism In Information Integration Based On Multi-Agent

With rapid development of Internet technology and information technology, multi-dimension information of different sources, formats or forms congests the network. How to accomplish uniform accessing these data requires integration of multi-dimension information. Now information integration based on multi-agent comes into being owing to the communication ability and cooperation ability of multi-agent system. However, it still exists some security problem as well as other multi-agent system, such as how to guarantee the security of host computer and the agent. Security is becoming the hotspot problem of the information integration base on multi-agent.When analysing the model of multi-agent information integration, security problem of that model is summarized, including host security, agent security and communication security. The strategy to resolve these problem is brought forth after that according to existing security technology. Then propose a security integration model based on Java security mechanism. This system makes use of existing Java security mechanism such as security manager, accessing controller, Java security packages. It realizes accessing controlling of the host resource, authentication of the agent, encryption of communication information and so on. Finally, it discusses the realization of security part of this system under the IBM Aglets platform.The security analysis of this system indicates that the security model can resolve most security problem of the integration system. It's safe enough for the demand of existing integration system. This model can make full use of the merit of Java language,and combine well with agent platform using Java. Researching on the security of information integration base on multi-agent can decrease the threaten of the security problem. It can push the actual use of integration platform, and it's referential to the traditional security research in information integration too.