Integrating mobile agents and swarm optimization for efficient QoS management in dynamic programmable networks

Mobile agents (MA) technology has been proposed for network and system management to solve the problem of scalability limitation of centralized models and the flexibility problems of static hierarchical networks. Mobile software agents represent a very attractive approach to the distributed control of computer networks and a valid alternative to the implementation of strategies for the management of QoS. Distributing intelligence across the network allows the fast exploration of more advanced services that can dynamically adapt to the user's requirements. This paper describes the efficient integration and adoption of mobile agents and swarm optimization in the implementation of an effective strategy for the development of efficient dynamic resource reservation by optimal rerouting (optimal alternative path allocation) under unpredicted dynamics which leads to an optimal migration strategy of tunnel agents in a constrained resource reservation model. We assume the RSVP protocol to be available at least in a subset of the network equipment, and investigate the use of swarm-based optimization with the problem of operating resource reservation in an IntServ model when some of the routers along the path do not support the RSVP protocol.. The proposed swarm intelligence optimization provides an adaptive, distributed searching process in order to identify the most effective resource allocation for QoS guarantee in large dynamic network between edge routers. We ran the experiments under various dynamics and the performance is compared with other routing algorithms. Our presented approach showed both very good performance and robustness under all the experimental conditions with respect to its competitors.