MagicNET: Security architecture for creation, classification, and validation of trusted mobile agents

Most of the current research and development results dealing with security of mobile agents describe solutions only for usage of mobile agents. These contributions usually assume agents posses unique and recognizable identities, cryptographic keys, assigned assurance level, and various other security parameters. But, very few papers describe how to create, classify, and evaluate mobile agents before their adoption and deployment. These are the issues we address in this paper: how mobile agents are created, validated, tested, and classified before their deployment. Based on certain security parameters, we establish classification scheme for mobile agents into three assurance levels: low, medium, and high. Our solution and procedures are based on a comprehensive security service-oriented architecture. We also describe roles, procedures, and security components used in that architecture. The most important results are the definition of assurance levels for mobile agents and procedures for their classification into those assurance levels.