Enhancing Response Selection in Impact Estimation Approaches

The number of attacks against computer systems is steadily increasing. Network administration personnel often have a wide variety of response measures against these attacks at their disposal. In previous work, a methodology was introduced for efficiently assessing the effects of countermeasures on network resources before their actual application and thus determining the most appropriate response. Building on this, we now propose a method of dynamically weighting the metrics used to evaluate the different responses. Instead of a fixed linear combination of metrics we introduce Pareto optimal combinations of the individual metrics and the combined cost measure. This allows a more flexible way of emphasizing the importance of individual metrics in different situations. The methodology was prototypically implemented in CoCoViLa, a powerful simulation engine for visually specified optimization problems.