Technique for policy conflict resolution using priority with variance

Abstract

A technique efficiently renders a policy-based decision from a plurality of options based on policy priority in a computer network. According to the novel technique, each of a plurality of policies is assigned an exclusive priority value, as well as a 'variance' specifying an acceptable value range (e.g., a percentage). Beginning at the highest priority policy, values corresponding to the plurality of options are analyzed based upon the current priority policy, and a 'best' option is determined according to the current policy. Thereafter, the variance for the current policy is applied to the value of the best option, defining a range of acceptable values for the current policy, and remaining options having an acceptable value for the current policy are established. If only one remaining option exists, that option is selected as the best option. Otherwise, the remaining options are analyzed based upon the next priority policy, a best option is determined, the variance is applied, and remaining options are established in the same manner, until a single best option exists.