ABSTRACT
Individual nodes are subject to a number of known security risks. Computer security systems identify entities (users or software agents) that access the system, determine the tasks each entity should be able to perform, and enforce policies restricting entities to performing only approved tasks [Tanenbaum 1997]. Designing systems that incorporate these concepts is challenging. Moreover, even a correct design will almost certainly have implementation flaws that introduce security risks into the system [Knight 1998]. These problems exist with all computer systems and are not unique to systems using mobile code; but, by allowing easier dissemination of software, mobile code heightens the risks.
