Strategic directions in database systems : Breaking out of the box

Database research must be more broadly defined than in the past. Database systems have developed in contexts where there is no control over the execution of their clients. This approach has led to a particular set of skills and techniques that can be applied and extended to other programs: 1. data modeling, 2. query languages, 3. query optimization and evaluation, 4. state-based views, 5. data management, 6. transactions, 7. distributed systems, and 8. scalable systems. Two applications of database technology that point the way for future data management systems are described: instant virtual enterprise and personal information systems. The technology to support these scenarios constitutes a research agenda for the next decade. A number of technical barriers that typically result from new application requirements not yet addressed or from the need for new database management systems architectures must be removed, including: 1. overhead, 2. schema organization, and 3. query complexity. In order to overcome these barriers a number of central research topics must be addressed, including: 1. imprecise results, 2. new transaction models, and 3. data movement.