Adaptive bandwidth control using fuzzy inference in policy-based network management

Abstract

This paper presents the fuzzy logic-based control structure for incoming traffic from an arbitrary node to provide admission control in a policy-based IP network management structure. The proposed control structure uses a scheme for deciding the network resource allocation depending on the requirement of predefined-policies and network states. The proposed scheme enhances policy adapting methods of existing binary methods, and can use resource of network more effectively to provide adaptive admission control, according to the unpredictable network states for predefined QoS policies. Simulation results show that the proposed controller improves the ratio of packet rejection up to 17%, because it performs the soft adaptation based on the network states instead of accept/reject actions in a conventional CAC (Connection Admission Controller). © Springer-Verlag Berlin Heidelberg 2005.