No one specific cause for glitches
We’ll have to wait and see if Nestor holds up
Ronald Stolk, director at the Centre for Information Technology (CIT), was upset when he heard about the malfunction that caused the cancellation of seven exams in November. ‘I can’t imagine how awful it is to be ready and prepared to sit an exam, only to have it be cancelled at the last moment’, he says.
‘But let’s remember that 96 percent of exams that day were administered. Although that didn’t help the 4 percent of students who couldn’t sit theirs.’ While the glitch followed two smaller ones that took place in October, Stolk did not see it coming.
‘We ran several tests over the summer with the online tools we’d be using this year, like the option to record classes and administer exams online’, says Stolk.
Stolk says a stress test, which involves checking whether the online environment has the capacity to handle the expected usage during the busiest times, was not performed, because usage during the summer is nothing like the volume of usage in November.
He emphasises that it’s not just about the system’s capacity – how many people work in Nestor simultaneously – but also about what people are using the system for. In a short period of time, Nestor transformed from an online environment that students would log onto quickly to check something, to an environment that is used often by thousands of students and lecturers every day for long periods of time, for both classes and exams. It turned out the system couldn’t cope with that change.
Not one specific cause
Stolk admits that he didn’t know what was going on during the November malfunction. ‘There wasn’t one specific cause we could point to. Together with the contractor we analysed the issue afterwards.’
The analysis showed the impact the difference in usage had on the way the Nestor system functioned.
Stolk says Nestor simply wasn’t equipped to handle the combination of the increased usage for online classes and the usage for online exams. But, he emphasises, there was no one specific cause for the glitch. ‘We can’t definitely point to something and say: that’s what went wrong.’
The analysis did lead to a possible solution: increase Nestor’s capacity. The capacity has also been adjusted to match the new way in which the system is being used. ‘If all goes well, we shouldn’t have any more glitches’, says Stolk.
The system is now being monitored for the issues that have come to light. However, Stolk would not be surprised if more glitches occur in the future. ‘As the past few months have proven, you never know.’