How to explain the fact that the alarm ports in the alarm list are fewer than the unqualified ports in the port qualification statistics as per China Telecommunication Standard in the DSLKeeper components?
1. In the statistic period, the alarm traps of some ADSL ports have not been switched on, or the alarm opening time falls within the statistic period. This causes the port alarm information during the time of period before switching on the port alarm to fail to be reported to the NMS. Therefore, the quantity of alarm ports is fewer than the quantity of unqualified ports.
2. Before the statistic period, some ports have been in the alarming status. The alarm status has not been recovered until the end of the statistic period. According to the port qualification rate standard, when a port generates an alarm upon non-conformance, it cannot recover the alarm until the non-conformance is removed. During this period, no alarm information is reported. Therefore, the NMS with such kind of ports cannot receive any alarm information. However, the statistics of unqualified ports is based on the data in connection parameters to judge whether the connection is qualified. So the alarm ports are fewer than the unqualified ports. It is a normal phenomenon.
The principle of statistics method for qualification of ports (during the statistic period):
Quantity of port connections = the union of the connections in the connection list and the alarms in the alarm list
The quantity of unqualified connections = the union of unqualified connections in the connection list and the alarms in the alarm list
The port qualification rate = (the quantity of port connections � the quantity of unqualified connections) / the quantity of port connections
If the port qualification rate = the threshold of qualification, then the port is qualified; otherwise, it is unqualified.
Therefore, it is normal that the alarm ports are fewer than the quantity of unqualified ports in the port qualification statistics as per the China Telecommunication Standard. It is not a problem of inaccurate statistics.