Divesh Srivastava

Data model and query evaluation in global information systems

Global information systems involve a large number of information sources distributed over computer networks. The variety of information sources and disparity of interfaces makes the task of easily locating and efficiently accessing information over the network very cumbersome. We describe an architecture for global information systems that is especially tailored to address the challenges raised in such an environment, and distinguish our architecture from architectures of multidatabase and distributed database systems. Our architecture is based on presenting a conceptually unified view of the information space to a user, specifying rich descriptions of the contents of the information sources, and using these descriptions for optimizing queries posed in the unified view. The contributions of this paper include: (1) we identify aspects of site descriptions that are useful in query optimization; (2) we describe query optimization techniques that minimize the number of information sources accessed; and (3) we demonstrate the need for interleaving planning and query execution in such a system, and present an algorithm for this purpose.

The CORAL deductive system

CORAL is a deductive system that supports a rich declarative language, and an interface to C++, which allows for a combination of declarative and imperative programming. A CORAL declarative program can be organized as a collection of interacting modules. CORAL supports a wide range of evaluation strategies, and automatically chooses an efficient strategy for each module in the program. Users can guide query optimization by selecting from a wide range of control choices. The CORAL system provides imperative constructs to update, insert, and delete facts. Users can program in a combination of declarative CORAL and C++ extended with CORAL primitives. A high degree of extensibility is provided by allowing C++ programmers to use the class structure of C++ to enhance the CORAL implementation. CORAL provides support for main-memory data and, using the EXODUS storage manager, disk-resident data. We present a comprehensive view of the system from broad design goals, the language, and the architecture, to language interfaces and implementation details.

Implementation of the CORAL deductive database system

CORAL is a deductive database system that supports a rich declarative language, provides a wide range of evaluation methods, and allows a combination of declarative and imperative programming. The data can be persistent on disk or can reside in main-memory. We describe the architecture and implementation of CORAL.nThere were two important goals in the design of the CORAL architecture: (1) to integrate the different evaluation strategies in a reasonable fashion, and (2) to allow users to influence the optimization techniques used so as to exploit the full power of the CORAL implementation. A CORAL declarative program can be organized as a collection of interacting modules and this modular structure is the key to satisfying both these goals. The high level module interface allows modules with different evaluation techniques to interact in a transparent fashion. Further, users can optionally tailor the execution of a program by selecting from among a wide range of control choices at the level of each module.nCORAL also has an interface with C++, and users can program in a combination of declarative CORAL, and C++ extended with CORAL primitives. A high degree of extensibility is provided by allowing C++ programmers to use the class structure of C++ to enhance the CORAL implementation.

Subsumption and indexing in constraint query languages with linear arithmetic constraints

Bottom-up evaluation of a program-query pair in a constraint query language (CQL) starts with the facts in the database and repeatedly applies the rules of the program, in iterations, to compute new facts, until we have reached a fixpoint. Checking if a fixpoint has been reached amounts to checking if any “new” facts were computed in an iteration. Such a check also enhances efficiency in that subsumed facts can be discarded, and not be used to make any further derivations in subsequent iterations, if we use Semi-naive evaluation. We show that the problem of subsumption in CQLs with linear arithmetic constraints is co-NP complete, and present a deterministic algorithm, based on the divide and conquer strategy, for this problem. We also identify polynomial-time sufficient conditions for subsumption and non-subsumption in CQLs with linear arithmetic constraints. We adapt indexing strategies from spatial databases for efficiently indexing facts in such a CQL: such indexing is crucial for performance in the presence of large databases. Based on a recent algorithm by C. Lassez and J.-L. Lassez for quantifier elimination, we present an incremental version of the algorithm to check for subsumption in CQLs with linear arithmetic constraints.

Rule ordering in bottom-up fixpoint evaluation of logic programs

Logic programs can be evaluated bottom-up by repeatedly applying all rules, in iterations, until the fixpoint is reached. However, it is often desirable-and, in some cases, e.g. programs with stratified negation, it is even necessary to guarantee the semantics-to apply the rules in some order. We present two algorithms that apply rules in a specified order without repeating inferences. One of them (GSN) is capable of dealing with a wide range of rule orderings, but with a little more overhead than the well-known seminaive algorithm (which we call BSN). The other (PSN) handles a smaller class of rule orderings, but with no overheads beyond those in BSN. We also demonstrate that by choosing a good ordering, we can reduce the number of rule applications (and thus the number of joins). We present a theoretical analysis of rule orderings and identify orderings that minimize the number of rule applications (for all possible instances of the base relations) with respect to a class of orderings called fair orderings. We also show that though nonfair orderings may do a little better on some data sets, they can do much worse on others. The analysis is supplemented by performance results. >

Bottom-up evaluation and query optimization of well-founded models

Abstract We present a bottom-up operational procedure for computing well-founded models of allowed programs with negation. This procedure provides a practical method of handling programs that involve unstratified negation in a manner that may be mixed with other evaluation approaches, such as semi-naive evaluation and various program transformations. We define classes of programs and sideways information of passing strategies (sips) for which the magic sets transformation preserves well-founded models with respect to the query. The classes of programs and sips we consider strictly subsume those already considered in the literature, and include stratified programs (with any choice of sips), left-to-right modularly stratified programs (with left-to-right sips) and arbitrary programs (with well-founded sips). For these programs and sips, our procedure for computing well-founded models is applicable to the rewritten programs, thus allowing increased efficiency of specializing a program for a query. Finally, we describe an extension of our bottom-up operational procedure that operates on the magic sets transformation of arbitrary programs with arbitrary sips, and computes query answers with respect to the well-founded models of the original program.

Pushing constraint selections

Recently, there have been attempts ([KKR90, Rev90, Cho90, BNW91] among others) to increase the expressive power of database query languages by integrating constraint paradigms with logic-based database query languages; such languages are referred to as constraint query languages (CQLs). Constraint programming paradigms are inherently declarative. Evaluating such programs can be expensive due to the manipulation of constraints, and hence optimizing such programs is very important. We consider the following problem:

Efficient Bottom-Up Evaluation of Logic Programs

In recent years, much work has been directed towards evaluating logic programs and queries on deductive databases by using an iterative bottom-up fixpoint computation. The resulting techniques offer an attractive alternative to Prolog-style top-down evaluation in several situations. They are sound and complete for positive Horn clause programs, are well-suited to applications with large volumes of data (facts), and can support a variety of extensions to the standard logic programming paradigm.

Explaining program execution in deductive systems

Programs in deductive database and programming systems have a natural meaning that is based upon their mathematical reading as logical rules. High-level ‘explanations’ of a program evaluation/execution can be constructed to provide added functionality: (1) To debug a program by following a chain of deductions leading to an unexpected (and possibly incorrect) conclusion; (2) To follow the derivation of certain correct conclusions to determine why and how they are reached; (3) To identify consequences of a (typically, incorrect or unexpected) fact. This functionality can be utilized either to perform post-mortem analysis of a session, or to interactively develop programs by running queries and viewing their deductions simultaneously.

The CORAL deductive database system

CORAL [4, 5] is a deductive database system that supports a powerful declarative query language. The language supports general Horn clause logic programs, extended with SQL-style groupiug, set-generation, aud negation. Programs can be organized into independently optimiied modules, and users can provide optimization hints in the form of high-level annotations. The system supports a wide variety of optimizw tion techniques. There is art interface to C++ that enables programs to be written in a combination of imperative and declarative styles; C++ code can be called from declarative programs, and vice versa. A notable feature of the CORAL system is that it is extensible. In particular, new data types can be defined, and new relation and index implementations can be added. An interface to the EXODUS storage manager [2] provides support for disk-resident data, transactions and crash-recovery.