IT infrastructure analysis, such as impact analysis, root cause analysis, and compliance checking, requires deep and detailed information about the dependencies between infrastructure components. Among the passive observation methods, flow-based network traffic information is a promising source to identify and label dependencies, although the so-called indirect dependencies are not directly observable from flow records. To identify indirect dependencies, network traffic analysis is needed. In this article a method is introduced that can discover context-sensitive indirect dependencies. The fundamental idea is to determine bipartite cliques and common interest groups of server pairs. From this information, a regular expression is made that can find indirect dependencies in flow records. After that the information can be propagated to CMDB to serve enhanced IT infrastructure analysis.
By: Istvan Szombath
Published in: RZ3714 in 2008
Questions about this service can be mailed to reports@us.ibm.com .
