Large Scale administration and information-gathering using CIM information modelling.
The normal miniCIM representation in AdCIM is semi-formal taxonomical classification that repesents domain knowledge, but is not entirely formal because many domain constraints and metadata are not expressed explicitly, so is not possible to infer and reason over the data without a priori knowledge of the semantics of the domain. In many cases, some assumptions about the domain are not explicitly coded, so. For example, in the domain of Wireless mesh networks, the channel information of a wireless interface actually maps into a range of frequencies that might be unusable because of national regulation or interference with other nearby equipment, so additional knowldege about frequencies and channels, their mapping and the problem of interfernce must be coded. The use of an ontology and a reasoner (a program implementing logical reasoning) allows to deduce a conflict in those situations.
Configuration semantic checking is another motivator, so constraints not only concerned with well-formedness, but with semantic integrity can be enforced. Problems in higher levels of abstraction are traced logically by the reasoner to lower-level causes and other configuration formats can be expressed with the same model, without changing the underlying logic.
We implemented a OWL transformation stylesheet in AdCIM that transforms schemas and instances to the OWL format (see here) and allows reasoning and policy application on AdCIM managed entitites. For more details, consult our publications.
