Improving Robustness, Flexibility, and Performance inDecision Support Systems

Companies often have to administrate a large amount of data. To make it more convenient and to get more useful information from that data, Decision Support System (DSS) are often used. The purpose of a DSS is to collect and analyse sets of data to facilitate decision making. One problem here is that there is often not a centralized system that handles all the data collection. This project researches the possibility of creating a centralized control system that can collect data from customers and set up new or modify configurations for different customers. There is also a problem when looking at performance. The centralized system must be able to handle large amounts of data transfers at the same time. To solve these problems a prototype of the system is implemented; this prototype is used to test both the auto-scaler and the configuration schema to see if they are sufficient for the task. With the help of the prototype, multiple tests are executed to test both the robustness and flexibility of the configuration schema and also to test the usefulness and the abilities of the auto-scaler. The results from all the tests are that the configuration in most cases works well enough for the task but there are still things to improve in the future. Looking at the results for the auto-scaler, the results show that the auto-scaler can proactively predict the number of transactions that will happen and scale in accordance with data. However, the auto-scaler is not able to keep performance at a constant level while the number of transfers increase. The conclusion of this project is that the current prototype, configuration schema and auto-scaler are a good start if wanting to centralize the administration of a DSS but there are still many things that can be improved to make the system even better and even more useful