ADFL Performance Boosting

The Challenge:

Very often we find ourselves in a situation where we have to extend or improve an already existing solution. In this project the client has been utilizing a custom built solution for trader mandates. The goal was to control what kind of products the different FO users are allowed to trade in a way more granular than what the standard Front Arena profiles allow you to do. The solution is using limits, query folders and text objects to store its data and has a process of granting and revoking mandates to users administered by the Middle Office. It has been in operation for years but its performance has always been quite slow. Administering the mandates which is done out of Operations Manager was particularly painfull. The following video shows the loading of the main workbook with all the mandates before the project. Unlike the users of this solution you don’t need to watch the whole video. Just skip to the end and check out the timer.

Our Improvements:

Since we had to work within the existing solution we had to take the data model as a given and find ways to improve performance on the Python and ADFL level. We first performed an analysis to find the bottlenecks in the performance. This solution does not have some heavy calculations so the performance issues come from loading the data. We identified redundant database calls and also designed more efficient ways to query the data and we were able to achieve a stunning 95% decrease of the loading time. Here is the same workbook loading after our improvements.