Thesis Research Project Dominik Paluch

Abstract: A rising number of enterprises migrate their enterprise applications at least partially to the cloud. Databases are an essential part of enterprise applications leading to a rising interest in database-as-a-service offerings. Virtualization and Multi-tenancy allows cloud providers to decrease hardware-, energy- and software licensing costs as well as reducing administration efforts. Thus, multi-tenancy has the potential to increase efficiency. However, predicting the performance of database tenants in different setups with varying workloads is a challenging task. Knowledge of the performance behavior is crucial in order to benefit from multi-tenancy. To enable performance simulations, we aim to provide automatically parametrized architecture-level performance models based on the Palladio Component Model for the in-memory database SAP HANA in a virtualized multi-tenant setup. These performance models contribute to practice by enabling cloud providers to evaluate the performance of their proposed virtualized multi-tenant setup before actually changing performance relevant parameters. Thus, they can avoid the violation of Service Level Agreements through specific measures.