Using a Statistical Approach for Security Parameter Determination

摘要

Determining appropriate security parameters is critical for most enterprises. An appropriately chosen parameter set will protect the enterprise against the threats impacting it while at the same time, minimizing their effect on systems performance. However, despite this being a critical need, rigorous methodologies have not been defined to solve this problem. A formal methodology can provide a sound basis for determining the optimal value of these parameters such that these tradeoffs are minimized. This paper demonstrates the application of statistical techniques for security parameter detection. Specifically, this paper demonstrates how fractional factorial analysis combined with ANOVA can be used to determine the value of theses parameters when the number of parameters under consideration is small. By way of illustrating this technique, this paper also determines the optimal values of network security parameters for a Solaris 9 platform hosting a critical production system.