Quality of service control in IP networks using Fuzzy Logic for policy condition evaluation

Abstract

This paper presents the architecture of a policy- based network management system designed specifically for Quality of Service management, where high level business policies are modeled as ECA (Event-Condition-Action) rules. Specifically, the system is mainly based on the policy architecture being proposed by the IETF (Internet Engineering Task Force), DMTF (Distributed Management Task Force), and TMF (TeleManagement Forum) standardization bodies. This work proposes a novel approach for policy condition evaluation using fuzzy logic. The fuzzy controller has the ability to examine policy conditions differently from default condition analyzers that employ simple conditions formed by a set of (IFs, ANDs and ORs), allowing the consideration and correlation of several input variables before taking decisions for the execution of policy actions. The system have been implemented and evaluated over a test bed network formed by Cisco® routers.