Future Directions of Database Research
- Changes in the VLDB Conference PC Structure -

At last year's VLDB Endowment meeting (New York, 1998), concerns were expressed that the area of database research may lose the pivotal role it now plays among information system technologies. It was agreed that the Endowment should maintain a watch on trends and future directions in the general area of information management to ensure that database research in general, and the VLDB conferences and journal in particular remain current and relevant. It was decided to set up a small "future directions" working group which maintains an ongoing dialogue with key researchers in the DB community and reports back to the Endowment Board. Working group members were Rakesh Agrawal, Michael Brodie, Michael Carey, Umesh Dayal, Jim Gray, Yannis Ioannidis, John Mylopoulos, Hans Schek, Kyu-Young Whang and Jennifer Widom.

The main observation of the working group was the following: While database technology has achieved a high standard in research and development, its future role in a globally distributed information network is less central. Considering the central role of databases in the past in providing a platform for application development, the members strongly felt that actions must be taken. Of course, it is true that databases still do a great job as storage managers for many applications. However, it was felt that the distance to applications and application development has grown substantially during the past decade. Among others, the following examples of application development and application areas were discussed as evidence for the need to act:

• Managing components, application services
• Distributed client/middleware/server computing
• Application frameworks, ERPs
• XML, e-commerce

To bring such topics into the fold of database conferences, the working group proposed to distinguish between the two main research directions:

1. core database technology
2. infrastructure for information system development

While (1) is well established and will be further developed worldwide by the international DB research and development community, the infrastructure direction (2) is under-developed and under-represented at conferences.

In all examples given above, the database role of providing a storage manager, i.e. core DB technology remains central. However, client application development takes place on other platforms provided by middleware technologies. Or client applications are obtained by customizing pre-fabricated ERP systems. The focus of our community should turn to the investigation of how core technology can become more widespread and usable, by concentrating on the description of new application areas, on the methods and tools for data analysis, design and integration, on the technologies for data deployment in modern architectures (middleware, wireless technology, the WEB), and in general on all the problems and challenges which are due to the need of using very large databases in new contexts.

In order to evolve into these directions the Endowment has decided during this year's Board meeting (Edinburgh, 1999) to take actions with regard to the selection of tutorials, panels and invited speakers for future VLDB conferences. Most important for the community is the following change in the VLDB Conference PC structure: While in the past there were two PCs for the scientific and industrial tracks tof the program, in future VLDB conferences a new track called "IS infrastructure and applications" will be added. Therefore in future Calls-for-Papers for VLDB conferences, three subcommittees will be distinguished, responsible respectively for papers on:

• Core DB technology
• IS infrastructure and applications
• Industrial applications and experience

The new strategy will be fully implemented in the year 2002 but efforts are undertaken to realize it earlier in the upcoming VLDB conferences in Cairo and Rome. For more information on the plans of the VLDB2000 organizing committee, please refer to the conference URL, http://www2.aucegypt.edu/vldb2000/.

Jim Gray, John Mylopoulos, Hans-J. Schek