Currently parallel object-relational database technology is setting the direction for the future of data management. A central enhancement of objectrelational database technology is the possibility to execute arbitrary user-defined functions within SQL statements. We show the limits of this approach and propose user-defined table operators as a new concept that allows the definition and implementation of arbitrary user-defined N-ary database operators, which can be programmed using SQL or Embedded SQL (with some extensions). Our approach leads to a new dimension of extensibility that allows to push more application code into the server with full support for efficient execution and parallel processing. Furthermore it allows performance enhancements of orders of magnitude for the evaluation of many queries with complex user-defined functions as we show for two concrete examples. Finally, our implementation perception guarantees that this approach fits well into the architectures of commercial object-relational database management systems.

We describe SECONDO, an extensible DBMS platform suitable for building research prototypes and for teaching architecture and implementation of database systems. It does not have a fixed data model, but is open for implementation of new models. SECONDO consists of three major components which can be used together or independently: (i) the kernel, which offers query processing over a set of implemented algebras, each offering some type constructors and operators, (ii) the optimizer, which implements the essential part of an SQL-like language, and (iii) the graphical user interface which is extensible by viewers for new data types and which provides a sophisticated viewer for spatial and spatio-temporal (moving) objects. Examples of algebras implemented in SECONDO are relations, spatial data types, R-trees, or midi objects (music files), each with suitable operations. The kernel is extensible by algebras, the optimizer by optimization rules and cost functions, and the GUI by viewers and display functions. A highlight of the description is a new algorithm for conjunctive query optimization which is remarkably simple, yet performs very well. We also emphasize a technique for selectivity estimation suitable for an extensible environment with complex algebras for non-standard data types.

This chapter is devoted to architectural and implementation aspects of spatiotemporal database management systems. It starts with a general introduction into architectures and commercial approaches to extending databases by spatiotemporal features. Thereafter, the prototype systems Concert, Secondo,

This document presents a method to design scalable and representative moving object data (MOD) and a set of queries for benchmarking spatio-temporal DBMS. Instead of programming a dedicated generator software, we use the existing Secondo DBMS to create benchmark data. The benchmark is based on a simulation scenario, where the positions of a sample of vehicles are observed for an arbitrary period of time within the street network of Berlin. We demonstrate the data generator’s extensibility by showing how to achieve more natural movement generation patterns, and how to disturb the vehicles ’ positions to create noisy data. As an application and for reference, we also present first benchmarking results for the Secondo DBMS. Such a benchmark is useful in several ways: It provides well-defined data sets and queries for experimental evaluations; it simplifies experimental repeatability; it emphasizes the development of complete systems; it points out weaknesses in existing systems motivating further research. Moreover, the BerlinMOD benchmark allows one to compare different representations of the same moving objects. 1

: We consider a heterogeneous database system (HDBMS) with a global object schema, and remote target databases that may be relational or object-oriented. The problem is to transform an initial query, posed against the global object schema in an object query language, into simplified queries that can be efficiently processed against some remote target databases. In the HDBMS, query optimization is made difficult by schematic discrepancy, and the need to model mapping information between the global and target schema(s). In this paper, we address this problem by representing the mappings from a target schema to the global schema, as a set of heterogeneous object equivalences, in an algebraic language. These equivalences are the basis for defining the heterogeneous search space of an HDBMS optimizer. We extend a modular rule-based query optimizer to produce simplified optimized queries for the HDBMS. The main result, reported in this paper, is the ability of the optimizer to uniformly per...

In this chapter we develop DBMS data models and query languages to deal with geometries changing over time. In contrast to most of the earlier work on this subject, these models and languages are capable of handling continuously changing geometries, or moving objects. We focus on two basic abstractions called moving

With the continuously advances in mobile devices, positioning technologies, geo-location wireless sensors or general data sensors, moving objects databases have been an important research subject in the last years. A comprehensive framework of abstract data types to describe objects that move freely or in networks has already been proposed. However, a moving object can produce additional data flows that describe the object state and that are of major interest for some applications like analyzing driving behavior or reconstructing the circumstances of an accident. In this paper, we analyze the limitations of a very well known algebra for moving objects in modeling the additional generated object data and propose an extension that accommodates the new requirements. Finally, we present our implementation strategy.

We study the representation and manipulation of geographic information in a database management system (DBMS). The conceptual geographic model that we use as a basis hinges on a complex object model, whose set and tuple constructors make it efficient for defining not only collections of geographic objects but also relationships between them. In addition, it allows easy manipulation of non-basic types such as spatial data types. We investigate the mapping of our reference model onto major commercial DBMS models, namely a relational model extended to abstract data types (ADT) and an object-oriented model. Our analysis shows the strengths and limits of the two model types for handling highly structured data with spatial components. Institut fur Informatik, Freie Universitat Berlin, D-14195 Berlin, Germany, voisard@inf.fu-berlin.de y Institut G'eographique National, 2 Avenue Pasteur, BP 68, F-94160 Saint-Mand'e, France Benoit.David@ign.fr 1 Introduction Many new applications that deal...