This sets the fault threshold, in microseconds, for average Inter-Frame Delay Variation, with the special rule (stated explicitly in the description) that a value of -1 disables fault generation for this metric entirely, while any value of 0 or greater is actively compared against the measured average IFDV to decide whether to raise a fault. An admin uses the -1 setting deliberately when average IFDV isn't a metric they want gating this service's fault status at all, rather than accidentally leaving a real numeric threshold in place that would be misapplied to a service where IFDV doesn't matter. For example, a bulk-data service not sensitive to inter-frame timing consistency might have cienaCesCfmDelayMsgAverageDelayVariationThreshold explicitly set to -1, while a voice or timing-sensitive service on the same switch keeps a real threshold like 30 actively enforced.