Highly Distributed Supervision for Autonomic Networks and Services

Abstract

Self-supervision and self-adaptability are key requirements for future network and service platform. This paper presents a fully distributed approach to the supervision of autonomic systems. The proposed approach relies on the self-adaptation features of autonomic elements and their clustering through self-organization algorithms. Furthermore, it presents two use cases of the proposed approach on fault handling and load balancing respectively. Moreover, the paper provides some initial evaluation, through simulation, on a load balancing use case. Finally, it compares the proposed approach with the related works.