Using Cyc(R) Technology to Create an Intelligent Network Configuration Tool

Investigate how to improve not just the precision and accuracy of networkconfiguration tools but also how to qualitatively expand the scope of whatsorts of conditions of interest they report, by incorporating a largecommonsense knowledge base to enable reasoning about network functionality. Current tools are limited by their inability to consider all the factorsthat bear on network configuration analysis, including adherence topolicies, network functionality, network design tradeoffs, and real-worldfactors (e.g., weather, holidays, physical layouts of offices, and so on).To make this happen, the network, the policies, etc. would have to bedeclaratively represented in a formal language such as CycL (predicatecalculus). Then a base of general knowledge similarly represented - e.g.the Cyc KB - could be brought to bear, along with the deductive reasoningengine associated with that KB. Altogether, this could be a power tool forthe network administrator, enabling them to detect multi-step conditions ofinterest (e.g., vulnerabilities) and to perform