Application-Oriented Security Policies and Their Composition

Abstract

Basically I seem to come back to this notion of policies once every five years when I remind myself of the logically difficult problem, namely that of determining the role of security administrators, if any, in systems in general. So the question is, can we shoot this guy and work happily everafter, and the answer is generally, I believe although I’m not quite sure, is that it is no. We basically have to have a certain degree of trust in the ability, intentions and motivations of security administrators. The second conclusion, which I think again is a partial conclusion, is that security administrators are really part of policies so in some sense we are examining the question about what should these good guys do as opposed to what the bad guys do. In other words, suppose that we solved the problem of the bad guys, there are no more bad guys in the world, no men in the middle attack, everything is fine, do we get security problems, and the answer is sure, human failures, bad intentions of this dependent component called a security administrator.