Self-Operation of a Network
Haitao Tang, Kaj Stenber · 2016
The current network operations are rule-based in a wide sense. In typical case the system and its functions are configured with inconsistent rules (or alike) to certain extent. Even if there does not exist inconsistency between the rules in some occasions, there is always a chance that the available rules cannot recommend what to do in unexpected situations which are not covered by the predefined rules. On the other hand, there are plenty of similarities existing between network elements, functions, and their relations. Based on the principles of case based reasoning, this paper provides a solution that can interact with a human operator and the functions of the system, can learn the experiences of the operations executed in the past, and can apply the learned operation experiences to solve new but similar problems. In this way the aforementioned inconsistency and incompleteness problems can be avoided. The architecture of this solution consists of a human user interface, a self-operation entity, and the functions of the system.