We have found the screening a case process involved the polling mechanism. Could you suggest the proper frequency of API call to poll the audit log for screened case ? Also, how about the average response time of whole screening process ? We may need the approximate time to evaluate our integration design.
For a deeper look into our World Check One API, look into:
408
●3 ●4 ●5
API Screening operation in MVP was designed to be an asynchronous(i.e. background) operation so it is not realistic to expect the operation can always be completed in almost real time.
The operation time will vary depending on the load of the system. When they can have the operation completed in a couple of seconds during periods of low load, the operation time can actually go over one minute during high load periods.
If you want to get results as soon as its ready, we think the best polling strategy that can work with such varying completion times would be to use exponential back-off (See https://en.wikipedia.org/wiki/Exponential_backoff ). For example, you can wait for 1 sec before making the first audit request and then can double the waiting time between each following call as follows: 2, 4, 8, 16, .... until you get you completion event.
408
●3 ●4 ●5
Yes that is correct. Below is the process/calls that you would ideally use to screen a name and get the results. Also please refer to documentation on sequence diagrams, which would help you to understand the flow in details further.
1) Save a name for screening
2) Queue a name for screening
3) Check for the audit history if the screening is ready (polling, would take couple of seconds)
4) Get the final screening results
On response times, it depends on several factors example name quality – proper full names, common/unique names.
