The Analysis Server performs root-cause and impact analysis within its specific OTM domain. It uses the Smarts patented Correlation Codebook Technology (CCT). The codebook contains an entry for each problem that can occur on each modeled object, together with the symptomatic events that result when that problem occurs. This unique list of symptoms is called the problem signature.
After initial discovery, and following topology updates to the repository, the codebook entries for the possible problems in the network are calculated by applying the problem definitions in the model to the actual network topology that has been discovered. This pre-calculation of problem signatures allows the Optical Transport Manager to quickly and accurately diagnose the root-cause problems in the network, using the currently active alarm conditions as input. When problems have overlapping signatures, the codebook algorithm proposes a set of most likely problems together with the probability that each one is a root cause. The Analysis Server compares the current event state of the network against the codebook at a user-configurable rate.
Equipment in optical networks is comprehensively instrumented in order to detect when a failure occurs, and to inform other network elements that a problem has occurred. This information is forwarded to management systems in the form of alarms.
A single fault in an optical network can result in many alarms being generated at different layers, such as section/regen, line/multiplexer, or path. Many of these alarms appear at points distant from the source of the problem. Using codebook correlation to perform root-cause analysis, the Optical Transport Manager Analysis Server identifies some of these alarms as symptomatic events of a root-cause problem, while it identifies other events as impacted by a root-cause problem.
