Adaptive Call Admission Control and Traffic-Policing for Multimedia Transmissions with Genetic Algorithms

摘要

In a network that carries multimedia transmissions, a new connection request (call) is admitted by Call Admission Control (CAC) if and only if the available network resources are adequate to satisfy its Quality of Service (QoS) requirements and it won't jeopardize the existing calls. During the call, traffic-policing entity monitors the traffic flow and drops the non-conformant packets. It's very desirable to make such an admission and policing process adaptive to the changes of the characters of the traffics and the network at the sources of traffics since these traffics are very bursty and highly unpredictable. The CAC and traffic-policing scheme we propose in this paper enables the network to work with end users based on flexible contracts and results in low average packet delay and packet dropping rates. It is built on and works as an extension to the "Cognitive Packet Network" (CPN) [18, 19, 20, 21] which is capable of dynamically choosing paths through a store and forward packet switching network like the Internet so as to provide best effort QoS to peer-to-peer connections. Genetic Algorithms (GA) are integrated into CPN [28,29] to help making smart routing decisions. Through analysis and simulations, we will show that CPN with GA works very well as a source routing platform for the proposed CAC and traffic-policing algorithms, as witnessed by the promising QoS measurement results presented in the end of the paper.