A rule engine for a computer network traverses a rule mesh having path nodes and path edges in form of a tree part and a graph part. The rule engine evaluates data packets flowing through a network to determine rules matched for every packet. Subsequent packets having same expression values as an already checked packet are not rechecked against the same nodes in the rule mesh through the use of a session entry. The rule engine performs a search on every path node of rule mesh to determine the next path edge to traverse. A Tree-Id and Rule Confirmation Bitmap that are indicative of path traversed and rules matched by a packet are generated at the end of rule mesh traversal. These are appended in the packet extension for subsequent modules of Policy Agent.