Ant Colony Optimization Based Self-Organizing QoS Framework in IP Networks

Abstract

We propose an ant colony optimization based self-organizing QoS framework, referred to as AntQoS. AntQoS introduces a new concept named QoS colony which is a virtual sub colony to heuristically search a sub-optimal path for given QoS requirements in real-time. Compared with previous QoS frameworks, AntQoS provides two promising capabilities using QoS colonies: 1) self-organizing network QoS (not pre-defined class-based QoS), which autonomously reconfigures itself by promptly adapting to the changing network status, and 2) self-organizing controls, which autonomously resolve network congestion and resist network failures. Simulation results show that AntQoS guarantees diverse QoS requirements of emerging services while increasing network throughput.