Correlating Variables and Performance Issues

In the case of dial, the MIB variables for cmLineStatisticsTable or the data gathered through the CLI show commands can and should be looked at and trended statistically. By looking at these variables over time, you can find trends, such as modems, that are connecting only at low speeds or that have more of a particular type of error. This type of analysis and correlation can help to diagnose performance problems within call groups, signaling levels, and even the lines. As you can see, a lot of what you gather in performance management can be applied to fault/error management.

The data gathered by show modem conn 56000 can be correlated with the data gathered from the RADIUS or AAA log. The use of the RADIUS or AAA log requires the use of several scripts and someone with a flair for programming. It initially is a great deal of work, requiring parsing of the RADIUS logs text fields, parsing the show modem conn fields, and populating a database with the results. At this point, you have a database against which the data can begin to be correlated with time and then reported on. This type of correlation between the two data sources can help validate the reports. In addition, the data from the RADIUS server can also provide a correlation point between the phone number and connection speed. The effort to build this type of correlation database can result in the ability to manage the dial modems and users for both performance and capacity planning.