Many transient conditions can affect the performance of a T1, if only for short periods of time. Effective monitoring requires that low-level details be suppressed to allow administrators to focus on solving high-level problems. ANSI T1.231 outlines one common approach, which underpins SNMP monitoring of T1 lines. A rough overview of the process is shown in Figure C-3.
Errors are the building blocks of the T1.231 approach.[34] An error is simply an unexpected transient condition. Although the specification forbids errors, they occur over short enough time spans that there is little performance impact from a single isolated error.
Defects are the result of persistent errors. Lone errors may be caused by any number of random events anywhere along the path of the T1, but defects persist for longer periods, so there may be a deterministic, resolvable cause. Defects may also lead to performance degradations.
Persistent defects lead to failures. Failures occur only when the line is down. No data can be transmitted, so immediate attention is required. Failures are pager material, but defects are usually not.
Performance data is the final component of the monitoring architecture. Carriers often guarantee certain levels of performance, but it is often up to the user to notice statistics falling below the guaranteed levels and demand a refund.
[34] New documents use the term anomaly instead of error. I find the former to be needlessly arcane, so I use the term error throughout this appendix.