Background: CORBA is a component architecture that enables inter-operability between software components that are based on heterogeneous platforms and in heterogeneous languages. CORBA goes further to specify a set of services providing low-level functionality to CORBA components. These CORBA services are vastly lacking in a specific area of functionality, this is accessing persistent data. The Persistent Object Service was the original service to provide this functionality, but has been discredited due to design faults and ambiguities resulting in the specification being withdrawn. This thesis presents an investigation into the analysis, design and implementation of a framework of services that together provide a persistent data access service. Objectives: The service should satisfy the missing functionality in CORBA, therefore opposing the need for developers to implement their own proprietary solutions. Proprietary solutions are expensive in developer resources and data access code is highly dependent on the proprietary data storage mechanism being used. The service will provide insulation from proprietary data storage mechanisms by encapsulating its data access interface, thus breaking the dependency of code on the specific data storage mechanism. The type of data and data storage mechanisms used in the Exploration & Production industry and their particular characteristics will be considered in the design of the service. Approach: The investigation has examined current methods of accessing persistent data from CORBA, including an in-depth analysis of the Persistent Object Service and a case study of a commercial application. The problems with these methods has been studied resulting in a list of requirements for the service to meet. A high-level design for the service has been outlined and a number of services have designed from bottom-up to reflect it. Results: The principal outcome of the research is the design of the Persistent Data Access Service that allows the manipulation of complex structured entity data. This data can be resident in any datastore, but has to be mapped to a standard data definition model. Other functionality that services that have been designed and implemented are providing distributed streams permitting copy-by-value and data transfer performance enhancement, distributed access to files and the transfer of complex structured entities in a stream. Implications: The Persistent Data Access Service gives to CORBA developers a standard means to access persistent data resident in heterogeneous datastores. This prevents applications becoming dependent on a specific data storage mechanisms and takes the responsibility away from the developer to implement their own proprietary solutions. The service also lets new CORBA applications and legacy applications run in parallel, as no changes has to be made to actual datastores and data models, unlike CORBA' s future Persistent State Service.