engleski

Self-organized Multi-agent System for Service Management in the Next Generation Networks

Next Generation Networks (NGN) aim to offer a wide variety of advanced telecom-munications and multimedia services. Introduction of these services will be enabled mainly by two factors: increased bandwidth in the access network and convergence of different legacy networks towards a universal all- IP core network. The offer of large number of services in the NGN environment will arise the need for service provision-ing and management procedures. Service management on emerging telecommunication systems that are distributed over a wide area is a hard task because it is not easy or even possible to perform final testing on a remote target system, as well as on system in operation [1-3]. Experiences show that it is possible for new software running on a target system to give a result different from the one obtained on test system [4]. The reasons are mostly the struc-tural and/or functional differences between both systems. Therefore, only implemen-tation and testing on the actual target system can give the answer whether the new software solves the problem (i.e., error, new operational circumstances, enhancement, and maintainability improvement) or not. Service management and software configu-ration operations in distributed systems become very demanding tasks as the number of computers and/or geographical distances between them grow. The situation gets worse with an increase in the complexity of the network and the number of nodes. This paper describes a method for service provisioning in an environment with a large number of distributed network servers and different versions of services placed across them. We have developed the agent based system called Remote Maintenance Shell (RMS) capable for remote control and management of services. To be specific, we present solutions for getting the service to the right place, starting it, and providing maintenance (upgrading with new versions). We have defined possible service distri-bution strategies for execution of operations and in this paper we consider applying of self-organized agents as a possible solution for this problem.

self-organized; multi-agent system; management; ngn

nije evidentirano