Object-relational Database Research Articles

Timer-driven database triggers and alerters

This paper proposes a simple model for a timer-driven triggering and alerting system. Such a system can be used with relational and object-relational databases systems. Timer-driven trigger systems have a number of advantages over traditional trigger systems that test trigger conditions and run trigger actions in response to update events. They are relatively easy to implement since they can be built using a middleware program that simply runs SQL statements against a DBMS. Also, they can check certain types of conditions, such as “a value did not change” or a “a value did not change by more than 10% in six months.” Such conditions may be of interest for a particular application, but cannot be checked correctly by an event-driven trigger system. Also, users may be perfectly happy being notified once a day, once a week, or even less often of certain conditions, depending on their application. Timer triggers are appropriate for these users. The semantics of timer triggers are defined here using a simple procedure. Timer triggers are meant to complement event-driven triggers, not replace them. We challenge the database research community to developed alternate algorithms and optimizations for processing timer triggers, provided that the semantics are the same as when using the simple procedure presented here.

Cost estimation of user-defined methods in object-relational database systems

In this paper we present a novel technique for cost estimation of user-defined methods in advanced database systems. This technique is based on multi-dimensional histograms. We explain how the system collects statistics on the method that a database user defines and adds to the system. From these statistics a multi-dimensional histogram is built. Afterwards, this histrogram can be used for estimating the cost of the target method whenever this method is referenced in a query. This cost estimation is needed by the optimizer of the database system since this cost estimation needs to know the cost of a method in order to place it at its optimal position in the Query Execution Plan (QEP). We explain here how our technique works and we provide an example to better verify its functionality.

Database patchwork on the Internet

Naturally, data processing requires three kinds of resources: the data itself, the functionality (i.e. database operations) and the machines on which to run the operations. Because of the Internet we believe that in the long run there will be alternative providers for all of these three resources for any given application. Data providers will bring more and more data and more and more different kinds of data to the net. Likewise, function providers will develop new methods to process and work with the data; e.g., function providers might develop new algorithms to compress data or to produce thumbnails out of large images and try to sell these on the Internet. It is also conceivable, that some people allow other people to use spare cycles of their idle machines in the Internet (as in the Condor system of the University of Wisconsin) or that some companies ( cycle providers ) even specialize on selling computing time to businesses that occasionally need to carry out very complex operations for which regular hardware is not sufficient. At the University of Passau, we are currently developing a distributed database system to be used in the Internet. The goal is to ultimately have a system which is able to run on any machine, manage any kind of data, import any kind of data from other systems and import any kind of database operations. The system is entirely written in Java. One of the most important features of the system is that it is capable of dynamically loading (external) query operators, written in Java and supplied by any function provider, and executing these query operators in concert with pre-defined and other external operators in order to evaluate a query. Compared to object-relational database systems, which allow to integrate external data and functionality by the means of extensions (datablades, extenders or cartridges) or heterogeneous database systems such as Garlic [MS97] or Tsimmis [GMPQ + 97], our approach makes it possible to place external query operators anywhere in a query evaluation plan as opposed to restricting the placement of external operations to the “access level” of plans. It would, for example, be possible to make our system execute a completely new relational join method, if somebody finds a new join method which is worth-while implementing. Because our system is written in Java, it is highly portable and could be used by data, function and cycle providers with almost no effort. Furthermore, our query engine is, of course, completely distributed providing all the required infrastructure for server-server communication, name services, etc.

OBJECT-RELATIONAL QUERIES INTO MULTIDIMENSIONAL DATABASES WITH THE ACTIVE DATA REPOSITORY

As computational power and storage capacity increase, processing and analyzing large volumes of multi-dimensional datasets play an increasingly important role in many domains of scientific research. Scientific applications that make use of very large scientific datasets have several important characteristics: datasets consist of complex data and are usually multi-dimensional; applications usually retrieve a subset of all the data available in the dataset; various application-specific operations are performed on the data items retrieved. Such applications can be supported by object-relational database management systems (OR-DBMSs). In addition to providing functionality to define new complex datatypes and user-defined functions, an OR-DBMS for scientific datasets should contain runtime support that will provide optimized storage for very large datasets and an execution environment for user-defined functions involving expensive operations. In this paper we describe an infrastructure, the Active Data Repository (ADR), which provides framework for building databases that enables integration of storage, retrieval and processing of multi-dimensional datasets on a parallel machine. The system architecture of ADR provides the functionality required from runtime support for an OR-DBMS that stores and processes scientific multi-dimensional datasets. We present the system architecture of the ADR, and experimental performance results for three applications implemented using ADR.

The BioImage Database Project: Organizing Multidimensional Biological Images in an Object-Relational Database

The BioImage Database Project collects and structures multidimensional data sets recorded by various microscopic techniques relevant to modern life sciences. It provides, as precisely as possible, the circumstances in which the sample was prepared and the data were recorded. It grants access to the actual data and maintains links between related data sets. In order to promote the interdisciplinary approach of modern science, it offers a large set of key words, which covers essentially all aspects of microscopy. Nonspecialists can, therefore, access and retrieve significant information recorded and submitted by specialists in other areas. A key issue of the undertaking is to exploit the available technology and to provide a well-defined yet flexible structure for dealing with data. Its pivotal element is, therefore, a modern object relational database that structures the metadata and ameliorates the provision of a complete service. The BioImage database can be accessed through the Internet.

SQL

For several years now, you've been hearing and reading about an emerging standard that everybody has been calling SQL3. Intended as a major enhancement of the current second generation SQL standard, commonly called SQL-92 because of the year it was published, SQL3 was originally planned to be issued in about 1996…but things didn't go as planned. As you may be aware, SQL3 has been characterized as “object-oriented SQL” and is the foundation for several object-relational database management systems (including Oracle's ORACLE8, Informix' Universal Server, IBM's DB2 Universal Database, and Cloudscape's Cloudscape, among others). This is widely viewed as a “good thing”, but it has had a downside, too: it took nearly 7 years to develop, instead of the planned 3 or 4. As we shall show, SQL:1999 is much more than merely SQL-92 plus object technology. It involves additional features that we consider to fall into SQL's relational heritage, as well as a total restructuring of the standards documents themselves with an eye towards more effective standards progression in the future.

Object management for a visual data analysis tool

This paper reports the advantages and disadvantages in applying an ODBMS (Object-oriented database management system) and an ORDBMS (Object relational database management system) to INFOVISER's object management scheme, a visual data analysis tool that we have been researching and developing. This paper also surveys a DBMS called LiteObject, which solves the main problems found in existing DBMSs, such as the complexity of implementing a versioning procedure and performance problems in loading a large number of instances. LiteObject offers version management based on constraints that exclude check in/out procedures from the application source code which decreases the degree of complexity. In addition, LiteObject offers a bulk loading interface to applications and manages several instances in one container. This reduces the amount of processing by removing identical steps in the process of inserting instances and adding instances to a collection type attribute. © 1999 John Wiley & Sons, Inc.

Object management for a visual data analysis tool

This paper reports the advantages and disadvantages in applying an ODBMS (Object-oriented database management system) and an ORDBMS (Object relational database management system) to INFOVISER's object management scheme, a visual data analysis tool that we have been researching and developing. This paper also surveys a DBMS called LiteObject, which solves the main problems found in existing DBMSs, such as the complexity of implementing a versioning procedure and performance problems in loading a large number of instances. LiteObject offers version management based on constraints that exclude check in/out procedures from the application source code which decreases the degree of complexity. In addition, LiteObject offers a bulk loading interface to applications and manages several instances in one container. This reduces the amount of processing by removing identical steps in the process of inserting instances and adding instances to a collection type attribute. © 1999 John Wiley & Sons, Inc.

B-SPID: an object-relational database architecture to store, retrieve, and manipulate neuroimaging data.

We propose a hardware and software architecture to respond to crucial problems in the neuroimaging field: storage, retrieval, and processing of large datasets. The B-SPID project, here discussed, concerns the processing of neuroimages and attached components stored in an object-relational multimedia database management system (DBMS). Advanced bioinformation concepts are exploited in this project such as large scale data storage, high level graphical user interfaces and 3D graphical processing and display of data. Our database implementation is based on standard programming components, runs on several UNIX platforms and is written to be evolutive. Queries on this database are designed to obtain and display from neuroimaging data several types of results (pictures, text, or 3D graphical shapes) on heterogeneous systems.

Extending the ODMG object model with composite objects

In this paper we extend the ODMG object data model with composite objects. A composite object is an object built by aggregating other component objects. Exclusiveness and dependency constraints, as well as referential integrity, can be associated with composition relationships among objects. Our composite object model is developed in the framework of the ODMG object database standard data model, but can be used in both object-oriented and object-relational database systems. In the paper, we propose a language for defining composite objects and we define the semantics of update operations on composite objects.

Enhanced abstract data types in object-relational databases

The explosion in complex multimedia content makes it crucial for database systems to support such data efficiently. This paper argues that the “blackbox” ADTs used in current object-relational systems inhibit their performance, thereby limiting their use in emerging applications. Instead, the next generation of object-relational database systems should be based on enhanced abstract data type (E-ADT) technology. An (E-ADT) can expose the semantics of its methods to the database system, thereby permitting advanced query optimizations. Fundamental architectural changes are required to build a database system with E-ADTs; the added functionality should not compromise the modularity of data types and the extensibility of the type system. The implementation issues have been explored through the development of E-ADTs in Predator. Initial performance results demonstrate an order of magnitude in performance improvements.

Efficient mid-query re-optimization of sub-optimal query execution plans

For a number of reasons, even the best query optimizers can very often produce sub-optimal query execution plans, leading to a significant degradation of performance. This is especially true in databases used for complex decision support queries and/or object-relational databases. In this paper, we describe an algorithm that detects sub-optimality of a query execution plan during query execution and attempts to correct the problem. The basic idea is to collect statistics at key points during the execution of a complex query. These statistics are then used to optimize the execution of the query, either by improving the resource allocation for that query, or by changing the execution plan for the remainder of the query. To ensure that this does not significantly slow down the normal execution of a query, the Query Optimizer carefully chooses what statistics to collect, when to collect them, and the circumstances under which to re-optimize the query. We describe an implementation of this algorithm in the Paradise Database System, and we report on performance studies, which indicate that this can result in significant improvements in the performance of complex queries.

Optimization techniques for queries with expensive methods

Object-relational database management systems allow knowledgeable users to define new data types as well as new methods (operators) for the types. This flexibility produces an attendant complexity, which must be handled in new ways for an object-relational database management system to be efficient. In this article we study techniques for optimizing queries that contain time-consuming methods. The focus of traditional query optimizers has been on the choice of join methods and orders; selections have been handled by “pushdown” rules. These rules apply selections in an arbitrary order before as many joins as possible, using th e assumption that selection takes no time. However, users of object-relational systems can embed complex methods in selections. Thus selections may take significant amounts of time, and the query optimization model must be enhanced. In this article we carefully define a query cost framework that incorporates both selectivity and cost estimates for selections. We develop an algorithm called Predicate Migration, and prove that it produces optimal plans for queries with expensive methods. We then describe our implementation of Predicate Migration in the commercial object-relational database management system Illustra, and discuss practical issues that affect our earlier assumptions. We compare Predicate Migration to a variety of simplier optimization techniques, and demonstrate that Predicate Migration is the best general solution to date. The alternative techniques we present may be useful for constrained workloads.

Secure and portable database extensibility

The functionality of extensible database servers can be augmented by user-defined functions (UDFs). However, the server's security and stability are concerns whenever new code is incorporated. Recently, there has been interest in the use of Java for database extensibility. This raises several questions: Does Java solve the security problems? How does it affect efficiency? We explore the tradeoffs involved in extending the PREDATOR object-relational database server using Java. We also describe some interesting details of our implementation. The issues examined in our study are security, efficiency, and portability. Our performance experiments compare Java-based extensibility with traditional alternatives in the native language of the server. We explore a variety of UDFs that differ in the amount of computation involved and in the quantity of data accessed. We also qualitatively compare the security and portability of the different alternatives. Our conclusion is that Java-based UDFs are a viable approach in terms of performance. However, there may be challenging design issues in integrating Java UDFs with existing database systems.

Database querying on the World Wide Web: UniGuide, an object-relational search engine for Australian universities

The World Wide Web can be considered to be a huge semi-structured database that can provide us with a vast amount of information. Existing Web search techniques have significant deficiencies with respect to robustness, flexibility and precision. The purpose of this research is to develop a domain-centred alternative to keyword and subject directory search engines. The specific domain being considered for the prototype implementation is that of Australian universities including all the internal entities that belong to each university such as faculties, departments, research centres, etc. that is available on the Web. By modelling the ontology of this particular domain using an object-relational data model and restructuring the Web data using an object-relational database, structured queries can be issued against this database in a fashion that current search engines do not provide.

PREDATOR

This paper describes PREDATOR, a freely available object-relational database system that has been developed at Cornell University. A major motivation in developing PREDATOR was to create a modern code base that could act as a research vehicle for the database community. Pursuing this goal, this paper briefly describes several features of the system that should make it attractive for database research and education.

A structured text ADT for object-relational databases

A structured text ADT for object-relational databases

Extending SGML to accommodate database functions: A methodological overview

A method for augmenting an SGML document repository with database functionality is presented. SGML (ISO 8879, 1986) has been widely accepted as a standard language for writing text with added structural information that gives the text greater applicability. Recently there has been a trend to use this structural information as meta-data in databases. The complex structure of documents, however, makes it difficult to directly map the structural information in documents to database structures. In particular, the flat nature of relational databases makes it extremely difficult to model documents that are inherently hierarchical in nature. Consequently, documents are modeled in object-oriented databases (Abite-boul, Cluet, & Milo, 1993), and object-relational databases (Holst, 1995), in which SGML documents are mapped into the corresponding database models and are later reconstructed as necessary. However, this mapping strategy is not natural and can potentially cause loss of information in the original SGML documents. Moreover, interfaces for building queries for current document databases are mostly built on form-based query techniques and do not use the “look and feel” of the documents. This article introduces an implementation method for a complex-object modeling technique specifically for SGML documents and describes interface techniques tailored for text databases. Some of the concepts for a Structured Document Database Management System (SDDBMS) specifically designed for SGML documents are described. A small survey of some current products is also presented to demonstrate the need for such a system. © 1997 John Wiley & Sons, Inc.

The BUCKY object-relational benchmark

According to various trade journals and corporate marketing machines, we are now on the verge of a revolution—the object-relational database revolution. Since we believe that no one should face a revolution without appropriate armaments, this paper presents BUCKY, a new benchmark for object-relational database systems. BUCKY is a query-oriented benchmark that tests many of the key features offered by object-relational systems, including row types and inheritance, references and path expressions, sets of atomic values and of references, methods and late binding, and user-defined abstract data types and their methods. To test the maturity of object-relational technology relative to relational technology, we provide both an object-relational version of BUCKY and a relational equivalent thereof (i.e., a relational BUCKY simulation). Finally, we briefly discuss the initial performance results and lessons that resulted from applying BUCKY to one of the early object-relational database system products.

Object-relational database systems (tutorial)

Object-relational database systems, a.k.a. “universal servers,” are emerging as the next major generation of commercial database system technology. Products from relational DBMS vendors including IBM, Informix, Oracle, UniSQL, and others, include object-relational features today, and all of the major vendors appear to be on course to delivering full object-relational support in their products over the next few years. In addition, the SQL3 standard is rapidly solidifying in this area. The goal of this tutorial is to explain what the key features are of object-relational database systems, review what today's products provide, and then look ahead to where these systems are heading. The presentation will be aimed at general SIGMOD audience, and should therefore be appropriate for users, practitioners, and/or researchers who want to learn about object-relational database systems.