Performance Evaluation of Replicated Data Management in the Cloud (Bachelor Thesis, Finished)


Fabio Tea


Clouds are very attractive environments for deploying different types of applications, from traditional online stores to

large scale applications, mainly due to their pay-as-you-go cost model and their highly available and scalable infrastructure.

Data management is an integral part of the applications deployed in the Cloud. Thus, it is of outmost importance

to provide highly available and scalable data management solutions. Data availability can be increased by using

well-known replication techniques. Different replication architectures and synchronization approaches

exist, from Master/Slave to Master/Master architectures combined with eager or lazy synchronization each

having different performance and monetary cost properties. Lazy synchronization has considerable advantages compared to eager synchronization when it comes to performance

and operational cost. However, it is well known that based on application characteristics lazy synchronization generates inconsistencies

(e.g. oversells) and thus may lead to high penalty costs.

The goal of this bachelor thesis is twofold. First, the development of tools for automated installation

and configuration of replicated IBM DB2 in the Cloud. Second, performance and cost evaluation of different DB2 replication approaches on the basis of the IBM Cloud environment.

Start / End Dates

2013/02/11 - 2013/06/10


Research Topics