Cloud computing has gained tremendous popularity in the past decade in the IT industry for its resource-sharing and cost-reducing nature. To move existing applications to the Cloud, they can be redesigned to fit into the Cloud paradigm, or migrate its existing components partially or totally to the Cloud. In application design, a three-tier architecture is often used, consisting of a presentation layer, a business logic layer, and a data layer. The presentation layer describes the interaction between application and user; the business layer provides the business logic; and the data layer deals with data storage. The data layer is further divided into the Data Access Layer which abstracts the data access functionality, and the Database Layer for data persistence and data manipulation.

In various occasions, corporations decide to move the their application’s database layer to the Cloud, due to the high resource consumption and maintenance cost. However, currently there is little support and guidance on how to enable appropriate data access to the Cloud. Moreover, the diversity and heterogeneity of database systems increase the difficulty of adaption for the existing presentation layer and business layer with the migrated database layer. In this thesis, we focus on the heterogeneity of SQL language across different database systems. We extend an existing open source Enterprise Service Bus with Cloud data access capability for the transformation of SQL statements used in the presentation and business layer, to the SQL dialect used in the Cloud database system back end. With the prototype we develop, we validate it against real world scenario with Cloud services, such as FlexiScale and Amazon RDS. Besides, we analyze the complexity of the algorithm we realized for parsing and transforming the SQL statements and prove the complexity through performance measurements.