Abstract

The growing popularity of cloud computing has led to an increasing demand for reliable evaluation and performance testing tools. In order to address this need, this thesis presents the development and evaluation of a workload generator bases on cloud computing patterns. The workload generator aims to replicate various workload patterns commonly observed in real-world cloud computing scenarios. The thesis also explores the concept, design strategy, and system requirements for the workload generator, outlining its key components and functionalities. The incorporation of specific and realistic modeling parameters helps achieve accurate representations of real-world behavior. While the workload generator successfully reproduces most workload patterns, certain discrepancies were identified, especially in capturing the essence of static workload patterns. Future work involves reassessing and refining the concept to improve accuracy and authenticity.