Towards achieving intelligent event correlation for end-to-end internet performance monitoring

The Internet has become a necessity to many organizations and also to the general public, similar to utilities such as electricity and water supply. Disruption or interruption to Internet availability would means reduced productivity and can be of negative consequence to many organizations. Therefore, it is critical that there is a mechanism to monitor Internet performance periodically in order to detect performance issues, and more importantly a mechanism to correlate the changes in the performance metrics with the event that causes the changes. Currently, there is no existing event correlation mechanism that is able to provide automated correlation between the changes in the end-to-end Internet performance with the occurrence of certain events or activities within an organization. Our goal is to develop an event correlation mechanism for end-to-end Internet performance monitoring which is automated and intelligent. In this paper, we describe in details the PingER monitoring framework, the challenges in interpreting the data collected from the framework, and finally discuss the requirements to develop an intelligent correlation engine for events correlation.