Using File Correlation to accelerate Decision-Making in a 
Decentralized Cooperative Security Enforcement


Abstract:
To enforce Chinese wall policies in a distributed file system 
decentralized cooperative security enforcement can be used. An example 
of such a security enforcement system is CliSeAu, which is developed 
at the MAIS group. For evaluation of CliSeAu in this setting a 
decision-making algorithm proposed by Gay, Hu, Mantel, and Sprick was 
used. One crucial aspect of this decision-making algorithm is that 
decisions are made whenever the system receives a request. As a 
result, decision-making overhead directly affects the system's 
performance. This motivates the idea to make decisions ahead of time 
to reduce the impact on the system's performance. 

In this talk we present the application setting we work with. We 
briefly give a high-level view on the current decision-making 
algorithm. Moreover, we present a vision of a new decision-making 
algorithm which makes decisions for likely occurring events ahead of 
time. Last but not least, we present our current design for the 
experimental evaluation of the two algorithms.