Cross Domain Correlation
For fault prediction in large technology infrastructures such as Telecom networks, Data Centers or in other environments where connected data flow is present, it is imperative that we can troubleshoot and conduct fault tracing in connected devices. This allows us to understand the hierarchy of faults essential to establishing a cause-effect relationship. The challenge is that in large networked setup many of these clusters need to be discovered automatically to extract generic cause-effect
Correlating faults automatically across different assets which can be from a different manufacturer, technology or domain has significant practical use in maintenance since it allows to group the related faults into a single ticket thereby making root cause analysis direct. Our foundational discoveries in large data stream clustering juxtaposed with operational setup provide an off-the-shelf application for fault cluster discovery.