Database Systems Supporting Next Decade’s Applications

Information is the master key to economic success and influence in the contemporary information society. “Only who can apply the newest information for his product development, is able to survive in the global competition” [Sch 95]. Crucial for the applicability of information is its quality and its fast availability. What is lacking most, however, is not the access to information resources but rather the facility to effectively and efficiently search for the required information. To cope with the information overkill will be the central competence of the next decade. Therefore, research on information systems will be one of the most important domains of computer science. To have connection with the worldwide leading group in this area will not only be of vital interest for academic scholars but even for the complete nationwide economy, in both branches traditional as well as the “new” economy. If the structure of the information to be searched is sufficiently simple, such as in one-dimensional numerical attributes or character strings, search problems can be considered as solved. Database management systems (DBMS) provide index structures for the management of such data [BM 77, Com 79] which are well-understood and widely applied. Requirements of traditional applications such as accounting and billing are perfectly met by a commercial DBMS. Therefore, the information infrastructure of most enterprises is based on products such as Oracle or Informix. In recent years, an increasing number of applications has emerged processing large amounts of complex, applicationspecific data objects [Jag 91, GM 93, FBF+ 94, FRM 94, ALSS 95, KSF+ 96]. In application domains such as multimedia, medical imaging, molecular biology, computer aided design, marketing and purchasing assistance, etc., a high efficiency of query processing is crucial due to the immense and even increasing size of current databases. The search in such databases, called non-standard databases, is rarely based on an exact match of objects. Instead, the search is often based on some notion of similarity which is specific to the application. For applications which do not only support transaction oriented search operations but also high-level decision making, it is necessary not only to search for objects which are similar to a given query object but rather to analyze the data set as a whole. Information which is interesting in the process of decision making are common patterns, classifications of data, knowledge about clusters (collections similar of objects), and, as the opposite of a cluster, exceptional data (outliers) which can be, for instance, indicators for the misuse of a system. This kind of information is commonly referred to as knowledge and the process of deriving such higher-level information from low-level transactional data is called data mining or (in the presence of a vast amount of data) knowledge discovery in databases (KDD). Because such applications on top of modern databases are also often based on some notion of similarity (or equivalently on the notion of data density) they also depend on similarity search. The difference to traditional similarity search applications is, however, that these applications do not only raise few, single similarity queries but rather a high number of such queries. From our point of view it will be more and more important for academic research as well as industrial practice to hold connection with the worldwide leaders in the corresponding areas of information science. It will be crucial to conduct internationally recognized research of high quality. We are planning to maintain and even strengthen our research which is oriented on publications in international top-level journals and conferences. Our main focus is on efficiency. Many problems of similarity search and data mining have been basically solved, and algorithms have been proposed that produce useful results. A general problem, however, are the vast amounts of data of today’s applications, and a processing time that makes the algorithms inoperative. Our general motivation is to change such