Load Testing the Programmes Plant

Staff in the Enrolment Management Services department continue to give feedback on the new course data administration system – most of it positive I am pleased to report. In the next few weeks we will carry out load testing on the system. We are currently working with staff who will use the system to create realistic scenarios to test against. This is never an easy task as many factors can influence how many people will be using the system at any one time – one of them being the design and usability of the application. Often there will be very few people logged into the system and even less actively using it but at certain peak times that number will be much higher. Working out how much higher is the problem. The best estimate we can make is that between twenty five and thirty would be a reasonable maximum to expect to be using the application during ‘normal’ periods. Not all of them will be processing data constantly – some may be using the system to check and locate data  – which will cause less strain on the servers. At peak times – say shortly before publication of a new prospectus – there may be as many as fifty or sixty users on the system.

 We are hoping for a positive result from the load testing – it certainly cracks along with only a couple of users on it. But if it does not perform as well as we would like there are things we can do and at this stage of the project there will still be time to do them. We could beef up the hardware and we can isolate processes or even lines of code that are causing bottlenecks. If these measures do not increase performance to an acceptable level we might have to look at staggered workflows.

 Our next technical sprint – aside from load testing – does not begin until November. We are using the time until then to look at how we might be able to re-use data which is managed and stored in other university systems such as the Student Data System and the Module Handbook Database. The ideal is one authoritative source of data but this is not so easy to achieve as one might first think. If data only flows in one direction – feeding the new Programmes Plant – then we will potentially have a disconnect with the source if data is edited in the new system. But we do not have the remit nor the resources to develop the originating application to keep it in synch.

Our current methods for storing, editing and publishing course data are circuitous and less than efficient with much potential for disconnects and differing versions of data. We have to face the fact that within this project we are unlikely to solve all the problems but if we design these first steps well we can build on this work and continue the good work in future projects.

 On the KIS data front we are ready to go though I think it is fair to say that the size and design of the KIS widget has not won it many fans around here. It would have been nice to  have had something that we could customize to tone with our online prospectus a little better.  Early days I guess…



Leave a Reply