SIDVI: a model for secure distributed data integration

The new millennium has brought about an increase in the use of business intelligence and knowledge management systems. The very foundations of these systems are the multitude of source databases that store the data. The ability to derive information from these databases is brought about by means of data integration. With the current emphasis on security in all walks of information and communication technology, a renewed interest must be placed in the systems that provide us with information; data integration systems. This dissertation investigates security issues at specific stages in the data integration cycle, with special reference to problems when performing data integration in a peer-topeer environment, as in distributed data integration. In the database environment we are concerned with the database itself and the media used to connect to and from the database. In distributed data integration, the concept of the database is redefined to the source database, from which we extract data and the storage database in which the integrated data is stored. This postulates three distinct areas in which to apply security, the data source, the network medium and the data store. All of these areas encompass data integration and must be considered holistically when implementing security. Data integration is never only one server or one database; it is various geographically dispersed components working together towards a common goal. It is important then that we consider all aspects involved when attempting to provide security for data integration. This dissertation will focus on the areas of security threats and investigates a model to ensure the integrity and security of data during the entire integration process. In order to ensure effective security in a data integration environment, that security, should be present at all stages, it should provide for end-to-end protection