Could you explain synchronous vs. asynchronous errored-seconds calculation?

問題：

Could you explain synchronous vs. asynchronous errored-seconds calculation?

答案：

For telecom applications, international practice is to use asynchronous errored seconds (ES). Synchronous ES is primarily used in North America. For digital video, Tektronix has probably set precedence by its use of asynchronous ES (first in the 1730D, in 1991, and in all models of the WFM601 series). Since digital video systems run at very low error rates, the difference between sync and asynchronous is probably academic. Synchronous vs. asynchronous errored seconds describes how error events are binned. In asynchronous ES, a clock is used to divide time into 1-second bins. When an error event occurs, it is put in its proper chronological bin. The errored second calculation simply looks at which bins have something in them (1 or more error events).

In synchronous ES, the bin is created when an error event occurs. The bin is made 1 sec long, so it will also collect any other error events that happen less than 1 sec after the initiating event. Note that the bin occurrence is synchronized with the error event, hence the name.

Since error events are associated differently between asynchronous and synchronous ES, the two methods can give different readings on the same signal. The difference will be greatest at moderate error levels. At very low error rates, errors are sparse and separated (assuming there is uniform distribution) and bins will only contain one error (In synchronous ES there will be few bins, each containing one error; in asynchronous there are many bins but the few non-empty bins will each contain one error. Thus the number of bins with errors will be the same, and the errored second reading will be the same). At very high error levels, there are one or more errors per second and so synchronous ES bin creation is continuous and it is essentially identical to asynchronous ES.

The advantage of synchronous ES is that two independent error monitors will read the same; their error gathering is synchronized by the error in the stream being monitored. Asynchronous monitors, since their time bases are independent, may group errors differently and hence give different ES readings. This will primarily occur at moderate error levels, as discussed above.