Dalia Motzkin

A stable quicksort

A sorting algorithm, called Stable Quicksort, is presented. the algorithm is comparable in speed with the Quicksort algorithm, but is stable. The experimental evidence presented support the theoretical evaluation of the performance of Stable Quicksort.

An efficient external sorting with minimal space requirement

An efficient external sorting algorithm with minimal space requirement is presented in this article. The average number of passes over the data is approximately 1 +Ln(N + 1)/4B, whereN is the number of records in the file to be sorted, andB is the buffer size. The external storage requirement is only the file itself, no additional disk space is required. The internal storage requirement is four buffers: two for input, and two for output. The buffer size can be adjusted to the available memory space. A stack of size log2N is also required.

Pracniques: Meansort

This paper presents an efficient algorithm based on Quicksort. The Quicksort algorithm is known to be one of the most efficient sorting techniques; however, one of the drawbacks of this method is its worst case situation of 0 (n2) comparisons. The algorithm presented here improves the average behavior of Quicksort and reduces the occurrence of worst case situations. 3~

An efficient directory system for document retrieval

This paper introduces a file directory structure which provides an efficient access path for document retrieval. The directory structure is based on the multi-B-tree structure. This directory structure is compatible with current automatic retrieval and query processing techniques. Weights that are assigned to index terms can be included in the directory with the terms at no additional cost . In addition, it provides for indexing a secondary attribute within a primary attribute with no additional cost . Updates are achieved with a high degree of efficiency as well. It is shown that this structure achieves a better overall performance than inverted files, standard B-trees, and other directory structures.

On high performance of updates within an efficient document retrieval system

Abstract This paper introduces fast, dynamic update algorithms for document retrieval systems. The M-B-T file directory structure, which provides an efficient access path for document retrieval, is utilized for the efficient updates. This directory structure was shown to be compatible with current automatic retrieval and query processing techniques, and to have the capability of including weights that are assigned to index terms and other secondary attribute values with no additional cost. Updates are achieved with a high degree of efficiency. It is shown that this structure achieves a better overall performance than inverted files, standard B-trees, and other directory structures.

On concurrent execution of information systems applications

This paper diaaaaaconcurrency problem that oeeurs during the execution of infonnaticm systems appliicms. The problem is not resolved by the systemsoftwaresuch as the underlyingdambaaemanagementsystemor the operadngsystem. Standardeoneumencycontrolmechanismsdo notaddressthispmbkm. Theproblemis identifiedas “thepmbkmofternporarycontaminationof data”.Whilecurrent coneummeycontrolmethodsaimtoachieveaeriakabilityanddurability, thesitu~atidh~~=mch ti~fli~~titil~ m~k~my~~e~ results. A easestudy where such incidents occurred is described Practical solutions as well as general thtmretical solutions are intro-

Uniform organization of inverted files

A range attribute is defined as an attribute that may assume a range of values. Examples might be Age = (1--10, 11--14, 15--16, ...) or Salary = (0-1000, 1001-1500, ...). This paper is concerned with the selection of ranges that will produce reasonably uniform numbers of records in each range. A set of algorithms has been developed to enable the file designer to obtain a set of ranges such that records are distributed uniformly between the ranges. Although in a given case perfect uniformity may not be achievable, the algorithms can find ranges such that for a set of X records in a range, bounds a and b may be given so that a ≤ X ≤ b for all ranges. The algorithms have been tested with a PASCAL program.

A tool for performance evaluation of database systems for small computer systems

This paper proposes a tool that provides a comparative evaluation of the performance of database systems. The tool has four components, the input component, the database-queries components, the statistical computation component and the report component. The input includes small, medium and large files as well as parameters relating to the database management systems being evaluated, the computers used for evaluation and ,the computer access normalization factor. The database-queries component incIudes several prepared sets of queries that test access paths, range retrievals, joins, query optimization and updates. The database management systems under evaluation executes the queries on small, medium and large files. The statistical computation component accrues the performance timings and executes statistical functions on these data. Finally, the report component provides the resulting evaluation in a tabular form. The tool has a user friendly interface. It provides reliable and accurate results even when the database management systems are running on different computers. INTRODUCTION One of the attractions of relational databases is that the user does not need to specify the access path to the data to be retrieved (Finkelstein, Schkolnick, & Tiberio 1988). Database management systems provide variety of indices that determine the paths (Mackert & Lohman 1986). Query optimization modules are also provided. Unfortunately, the database management system manuals contain little or no information on performance. Furthermore, indexing does not always perform according to specifications. The literature provides a considerable number of theoretical analyses of performance issues (Bitton and Turbyiill, 1988; Ip et al., 1983; Mackert and Lohman,1989, Motzkin, 1991 and many others). However, very little is available about the performance of commercial database systems. Some information is available on the performance of SQL (Ortali, 1988), and INGRES (Rowe and Stonebraker, 1986; Youssefi and Wong, 1986). Information may be available in the form of internal memos or reports in some corporations. Unfortunately the user has no easy way of knowing what performance to expect. There “Permission to copy without fee alI or part of this material is granted provided that the copies are not made or distributed for direct commerical advantage, the ACM copyright notice and the title of the publication and its date appear, and notice is given that copying is by Penmssion of the Association for Computing Machinery. To copy otherwise, or to republish, requires a fee and/or specific permission.” @ 19% ACM O-89791-658-1 95 0002 3.50 are no convenient tools that perform evaluation and comparison of database management system. The tool proposed here provides performance evaluation of database systems. It determines the conditions under which indexes perform the best, as well as those under which they do not perform well. The input includes small, medium and large files, information relating to the database management systems being evaluated, the computers used for evaluation and the computer acc:ess normalization factor. The tool includes several sets of database queries that test access paths, range retrievals, joins, query optimization and updates. The statistical computation component of the tool accrues the performance timings and executes statistical functions on these data. Finally, the report component provides the resulting evaluation in a tabular form. The performance evaluation tool PET has a user friendly interface. It is very easy to use. The user needs only to load the database system into the tool using the Create and Open Database buttons, create indexes if so desired and click the analysis button. The pre-stored queries are then executed followed by statistical analysis. Reliable and accurate results in report form are then displayed. The proposed tool is based on an early performance model developed by Motzkin & Boume (Motzkin & Boume 1993). THE PERFORMANCE EVALUATION TOOL In this section we describe the Performance Evaluation Tool PET that we designed and implemented. The PET allows the user to create the database, load the data from a file, create the indices, drop the indices, run groups of queries , set the configuration of the system, analyze the timings and generate on screen and hard copy results. The PET provides on-line help for each of its functions. The graphical user interface is provided using VISUAL BASIC. The current version of the PET runs under Microsoft Windows. The overall structure When a user initiates PET, he/she is presented with the Main Menu of the system. The main menu is as shown in Figure 1. Each function of the menu can be invoked by clicking the appropriate command button. As will be seen in the following sections each of the functions has its own set of windows, when the user is done with the tool, he/she returns to the main menu

Distributed database design—optimization vs feasibility

Abstract This paper is concerned with the problem of optimal assignment of data to sites in a distributed relational database. It is shown that in general the optimal allocation will require exponential time in terms of the input. Several heuristic algorithms that can be applied to various constraints and which provide feasible, near optimal results, as well as a model that determines the “best” assignment for a given input out of several optimal and near optimal assignments have been developed. The model is shown to be efficient, to require polynomial time, to be practical in terms of feasible inputs and to achieve assignments with near minimal global and local costs.

Parallel organization and performance of an information system

This paper investigates a directory structure, based on an M-B-T directory, that provides for parallel retrieval of information with low communication overhead. An information system of documents is assumed. It is compatible with current automatic retrieval and query processing techniques. It is also compatible with any organization of the documents. Weights that are assigned to keywords can be included in the directory with the keywords at no additional cost. In addition, it provides for indexing a secondary keyword within a primary keyword with no additional cost. Given a p-processor massively parallel computer, our organization allows a maximum of p retrievals to be performed concurrently and thus achieves an approximate speed up of p in the best case. The new, revised organization improves on previous work. An eecient parallel searching algorithm and its performance is also discussed.