Abstract
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.
| Original language | English |
|---|---|
| Pages (from-to) | 727-739 |
| Journal | Engineering Applications of Artificial Intelligence |
| Volume | 17 |
| Issue number | 7 |
| DOIs | |
| Publication status | Published (in print/issue) - 1 Oct 2004 |
Bibliographical note
Other Details------------------------------------
This paper is significant because it (i) introduced a novel human-computer collaborative discovery methodology and associated tools for acquiring domain knowledge (in the form of correlation rules and beliefs) from complex communications networks, to facilitate self-management; and (ii) describes a process that evolves to be increasingly autonomic/self-discovering, while providing drill down and visibility for the human. The event correlation research in this paper was undertaken in a collaborative project with Nortel Networks (1999 – 2002) and its extension to autonomic communications was explored through a British Telecom (BT Exact) Short Term Research Fellowship (2003).
Keywords
- Keywords: Autonomic networks
- Autonomic communications
- Problem determination
- Root cause analysis
- Self-healing
- Event correlation
- Development framework
- Development tools