Workflow systems provide automated support that enables organizations to efficiently and reliably move important data through their routine business processes. For some organizations, the information processed by their workflow systems is highly valued and in need of protection from disclosure or corruption. Current workflow systems do not help organizations to adequately protect this important data. We describe a prototype secure workflow system that allows users to develop high level workflow security policies and to automatically execute these policies within the workflow system. These workflow policies can use the workflow context to provide fine-grained dynamic access control and other security services that enhance the organization's ability to control the information contained in its workflow system. We explain these security policy goals, our prototype policy editor, our prototype workflow server, and our underlying Java based implementation.

A representation for concurrent programs, called task inter-action graphs, is presented. Task interaction graphs divide a program into maximal sequential regions connected by edges rep-resenting task interactions. This representation is illustrated and it is shown how it can be used to create concurrency graph rep-resentations that are much smaller than those created from con-trol flow graph representations. Both task interaction graphs and their corresponding concurrency graphs facilitate analysis of concurrent programs. Some analyses and optimizations on these representations are also described. 1

The main result of this paper is that obtaining any information about the $O(\log |p|)$ “most significant” bits of x, given $g^x (\bmod p)$, even with a tiny advantage over guessing, is equivalent to computing discrete logarithms $\bmod p$.

Blum and Micali [4] showed how to hide one bit using the discrete logarithm function. In this paper we show how to hide c•loglog p bits for any constant c, where p is the modulus.