Implementation and evaluation of a cloud solution as an optimization problem of the performance of an on-premise web application.

The present work is based on a real business case, supporting a company to optimize the performance of their on-premise e-commerce platform by proposing a cloud solution, running performance testing and confronting the achieved results. For the company it is essential to maintain fast server response times that do not fluctuate due to the increased user activity. The existing standard solutions that address this problem intuitively suggest adding more processing power, however this would increase the upfront cost which is not an efficient choice in the long term. Our interpretation is that a cloud solution would minimize the workload and optimize the resource usage to increase throughput and minimize contention of the database, that drastically effects the performance of a web application. The conducted activities in this project involve gathering business, technical, and user requirements, compiling a requirements definition, performing baseline measurements, designing and implementing a cloud solution on the AWS platform, and finally repeating the tests to measure the performance improvement. The solution is evaluated following a statistical analysis approach. The results of the empirical tests are based on a set of computational metrics that give quantitative estimations to evaluate the effectiveness of the system. The comparison between the baseline measurements and the implemented cloud solution prototype, demonstrated a significant optimization of throughput and server response time.