Autonomic networks: engineering the self-healing property

As the size and complexity of networks and communications continue to grow, there is a heightened need to develop new techniques capable of achieving a level of service with successful operations upon which users can place even more reliance. Key emerging strategies for meeting this demand is 'autonomic networks' and 'autonomic communications', concepts similar to autonomic computing while specific to the communications field. This paper considers the 'self-healing' aspect of autonomic networks, examining, in particular, techniques for event correlation to aid fault identification. A three-tier rule-discovery framework and associated support and analysis tools are described. These assist with the development, management and maintenance of correlation rules and beliefs. (C) 2004 Elsevier Ltd. All rights reserved.