The integration of mobile agents with SNMP creates significant advantages for the management of complex networks. Nevertheless, the security concerns of mobile agent technology limit its acceptance in practice. A key issue is to safeguard resource usage abuse by malicious or buggy mobile agents on the hosting system. This paper describes how the TINMAN architecture, a framework and a suite of tools for enforcing resource safety of mobile code is applied to mobile SNMP agents. TINMAN uses a suite of resource-usage checking tools which consists of a resource bound predictor, a usage certification generator and a verifier at compile-time, and certificate validation and monitoring tools at run-time. This paper shows how TINMAN tools can provide 100% coverage by a combination of off-line static analysis and run-time monitoring in enforcing safety on resource consumption of mobile SNMP agents. Experimental results from the current TINMAN implementation are given.
Weijiang Yu, Aloysius K. Mok