Paul Pester has now spoken and from what he said it appears that the back end was OK. but that the customer facing or perhaps interfacing part had issues.
Quote, ‘Testing before the switch were “misleading” as they did not foresee the problems with the system’s capacity.’
From my experience of cutovers throughput capacity estimation, including the first day stimulus factor (never to be underestimated) and testing are a serious issue. Not least because generating test traffic is a pain and estimating the choke methods needed to manage or constrain ‘first day access’ loads appears to be another black art. It is simply not easy when you get 100% of normal total daily demand in less than an hour and I have seen that happen. Repeated access attempts then add to the load as erroneous inputs increase.
.
