A Multi-Agent Flexible Architecture for Autonomic Services and Network Management

Abstract

Modern data and communication networks have been growing dramatically in diversity, heterogeneity and complexity. Hence the management of such dynamic networks is an extremely difficult task due to some key issues such as; the increasing network size and functionality, rapidly and continuously changing network topology and complexity. This paper introduces a management paradigm to give rise to the autonomic management solutions (i.e., activation, monitoring, control, and administration) of management functionalities. It also introduces an architecture that (1) places the various management functions (FCAPS) into a design structure that can be used to select those functions that may give way to autonomic management and (2) to design the algorithms that may fulfill the management function. This architecture consists of four horizontal logical layers (Products, Components, Services, and Resources Layers) and two vertical logical layers (Management layer and information layer). Autonomous management agents (Swarm agents in this context) sit in the management layer and can sense changes/events in the four horizontal layers and react autonomically to reconfigure and adapt themselves based on the information scattered and gathered over the network.