UNIX system security mostly presumes that a computer's backplane can be trusted because computing operations are assumed to be local, and because the computer itself can be physically secured. In a distributed environment, the logical equivalent of the single system's backplane is the network itself. Network computing means distributed, rather than localized, computing operations and, in the case of an open network (which DCE assumes), little of the network is physically secure. Thus, the nature of distributed systems poses special security risks, in addition to those posed by nondistributed systems. Unlike UNIX system security, DCE security is designed specifically to address those risks.
These considerations notwithstanding, network security is ultimately dependent on the security features that are local to the individual computers in the network and, what is more important, the manner in which those features are used and administered. Since any compromise to the local security of a computer in the distributed environment may introduce opportunities for compromising network security, DCE security does not diminish the importance of local security. In fact, the relative importance of local system security is greater in the distributed environment because the consequences of a local security breach may not be local. Finally, while DCE security does nothing to enhance local security, neither does it introduce any new avenues for compromising local security.
In the discussions in this guide, we assume you are familiar with the authentication and authorization features that UNIX systems provide: /etc/passwd and /etc/group file processing, routines that return or change file attributes, routines that return or change real or effective user IDs (UIDs) and group IDs (GIDs), and data encryption and decryption.