UTT-based Outage Detection

The CNA system maintains a rolling average of the time interval between measurements for each prefix through each service provider link. (This interval is displayed as IMT, or inter-measurement time, in the show prefix measurements command output.)

A significant increase in the time between measurements could signify an outage in service (sometimes called a black hole) for a particular prefix across a particular link.

If the increase is unacceptable according to user-defined limits, the CNA system will consider it an outage and, if the link is the currently asserted winner, the system will immediately recompute the performance route for that prefix.

This behavior is controlled by the outage-detection commands:

• outage-detection silence--determines when an outage has occurred on a single link.
• outage-detection total-silence--determines when an outage has occurred over all of your links
• outage-detection active-probe--Defines the number of consecutive active-probe failures that should be interpreted by the CNA decision-maker process as an outage for the prefix.
• outage-detection brownout--Defines the parameters used by CNA for detecting network brownout conditions.

Use the show statistics dmaker command (or view engine statistics in the web interface) to see the current settings.