Michael E. Senko

Data structures and accessing in data-base systems: I evolution of information systems

The attention of the computer user community is increasingly focusing on data bases and computerized information systems because of two converging trends. Computerized information systems are coming to play an essential role in business operations, and the hardware and software technology for supporting information systems is in a period of rapid technical progress. In spite of this attention, there is still little common agreement as to what information systems are, the functions they perform, and — from a technical point of view -how they should be designed, implemented, installed, and used. These conditions arise naturally from the newness of the field and the ad hoc nature of existing Computerized implementations. A compounding factor is the wide variety of perspectives of those who discuss information systems — for example, the executive, the management consultant, the systems analyst, the mathematician, and the systems programmer. In this paper, we emphasize aspects of information systems that are of importance to systems analysts, systems designers, and implementors.

Information systems: Records, relations, sets, entities, and things

Abstract This article reviews progress in the creation of a scientific discipline for information systems. It discusses contributions from four sources of fundamental knowledge: 1. (1) Information Systems Technology; 2. (2) Scientific Computation Technology; 3. (3) Linguistics; and 4. (4) Mathematics. It then selectively reviews progress on an information systems science in the most active areas of study: name-based representations, stored representations, access languages, and information systems performance. This discussion relys on the definition of a series of abstract, structured levels for the description, design, and implementation of generalized data base management systems.

File Organization: On the Selection of Random Access Index Points for Sequential Files

The construction of a hierarchy of indexes (the indexed sequential access method) is one means of providing rapid random access to sequential files. An examination is made of the consequences of partially or completely replacing one or more index levels by linear interpolation procedures. For all possible configurations of the several types of key distributions investigated, linear interpolation on the average provides significant performance improvements. Typically, the two accesses required to obtain track index and data are reduced to 1.1 to 1.7 accesses per record. Extremely unusual key distribution will, however, raise the number of accesses required above 2.

Specification of stored data structures and desired output results in DIAM II with FORAL

The primary goal for a specification language is to provide a means for accurate, succinct description of processing algorithms and their related data structures: in effect, a description with a minimum of extraneous detail. In general, progress toward this goal has been achieved by moving from assembly languages with their requirement for detailed knowledge of data processing structures to notations, such as the mathematical formalisms, which avoid the use of specific computer representations. In this sense, FORTRAN, ALGOL, and PL/I are early specification languages, while APL, the Vienna Definition Language, and SETL are more advanced forms.

Data structures and data accessing in data base systems past, present, future

A broad range of commercial and research data base systems are analyzed. Common characteristics are discussed. These systems, which have roots in older filing systems and in punched card systems, are grouped into the three categories of hierarchic, network, and single-level models. Also presented is work on the standardization of data base systems. Research toward the discovery of new commonalities is also discussed. This paper is based on an extensive published literature.

Concepts of a Data Independent Accessing Model

The Data Independent Accessing Model (DIAM) represents a constructive combination of the formal mathematical and the evolutionary pragmatic approaches to a Generalized Data Base Management System. It aims at a complete separation between the users (application programmers) model of real-world objects, events, and the relationships among them, and the internal representation of those objects, events, and relationships. At the same time, it aims at complete flexibility of internal representation in terms of simple specifications, the effects of which are easy to understand. The DIAM encompasses four levels of descriptive models (Figure 1), each providing for flexible augmentation of the descriptions of the preceding level. Specifications at each level are retained in a catalog.

A cost oriented algorithm for data set allocation in storage hierarchies

Data set allocation in todays multilevel storage systems is usually based on qualitative, ad hoc decisions. While it would be desirable to obtain an optimal solution to this allocation problem, it is clear that the number of parameters involved makes it intractable to straight-forward solution. In such a situation, we must find a set of assumptions which simplify the problem greatly, but which still provide a basis for considering all significant cost elements. This paper presents such a first, quantitative allocation step. It considers many of the significant detailed costs of system utilization, data storage, data staging, and data migration. Although many avenues of further improvement are available, the present algorithm seems to be usefully accurate. As such, it can aid in quantifying the problems of data set allocation, storage system configuration, and new device designs.

Analysis of a complex data management access method by simulation modeling

The typical paper on file organization presents a qualitative discussion of a proposed search structure. In a very few instances, the writer may include one or two numbers to indicate system performance. In reading these papers, one is led to the impression that file design is an extremely simple problem and that intuitive guesses can easily provide optimal answers. However, the sophisticated file designer knows from experience that there are many relevant parameters and that they can drastically affect performance. This knowledge can, however, be frustrating because he cannot afford the machine time and resources to obtain a good design by performing an adequate parametric study on possible file organizations. (For example, setting up a large file in just one configuration can require several hours of machine time and several man days or weeks.)

Specifications in a Data Independent Accessing Model

The Data Independent Accessing Model (DIAM) Project of the Information Sciences Department of IBM Research is directed towards developing an architectural basis for an advanced data base system - a data base system which, in addition to providing advanced functional capabilities and a new level of data independence, is not limited either in the access paths which can be declared (in order to support the user-specified information collections) or in the encodings which are possible for those access paths. We have already been presented with a general overview of the basic system (1). In that overview, DIAM was seen to be composed of 4 basic levels of description: the Entity Set Model the String Structure Model the Encoding Model the Physical Device Model. In this paper we shall concentrate on the middle two of these levels - the String Structure and the Encoding Models.

A query-maintenance language for the data independent accessing model II

Abstract FORAL is a data base language designed specifically for the access to binary semantic network structures at the DIAM II Infological Level. It is a non-procedural language, that has some of the characteristics of natural English. The user writes transactions in terms of real world things and their attributes rather than in terms of fields, records, and files. A review of the original FORAL led to FORAL II described in this paper, offering a more readable syntax, in particular for long documents.