Sushil Jajodia

Managing replicated files in partitioned distributed database systems

In this paper, we describe a consistency control algorithm for managing replicated files in the face of network partitioning due to node or communication link failures. It adopts a conservative approach in that mutual consistency among copies of a file is maintained by permitting files to be accessed only in a single partition. Our algorithm has the property that it permits dynamic switching between the “dynamic voting” algorithm and the “linearly ordered copies” algorithm. This aspect is not only appealing but results in greater file availability than in all previously published conservative algorithms.

Alternative methods for the reconstruction of trees from their traversals

It is well-known that given the inorder traversal of a binary trees nodes, along with either one of its preorder or postorder traversals, the original binary tree can be reconstructed using a recursive algorithm. In this short note we provide a short, elegent, iterative solution to this classical problem.

Translation of entity-relationship diagrams into relational structures

Abstract An entity-relationship diagram (ERD) is considered to have two conceptual domains: its upper and lower conceptual domains. The semantic properties intrinsic in the upper conceptual domain were studied recently by Lien. In this paper, we discuss the properties inherent in the lower conceptual domain. We show that the latter contains a significant portion of the semantic information of an ERD and helps determine its data representation power. We develop the notions of fundamental functional dependencies (FDs) which can be readily recognized for the lower conceptual domain of an ERD, and regular ERDs which are ERDs such that all the FDs present among their attributes can be derived from their fundamental FDs. Then we give conditions under which a large number of ERDs are equivalent to their underlying relational structures.

Entity-relationship diagrams which are in BCNF

In Ref. 8, we introduced a simplifying assumption about entity-relationship diagrams (ERDs), called regularity, and showed that regular ERDs have several desirable properties. One such property is that every relation schema in the ERDs canonical relational scheme can be put into Third Normal Form. We left open there the more basic question: under what conditions would the original relation schemas actually be in Boyce-Codd Normal Form (BCNF)? Since the visible semantics of ERDs determine naturally their associated functional dependencies (fds), it is important to know when an ERD, as designed, already has this strongest normal form given purely in terms of fds. We show here a sufficient diagrammatic condition (loop-free) under which a regular ERD will have databases enjoying the benefits of BCNF.

Surjective stability of units and simple homotopy type

Recently M.N. Dyer has shown that the classification of CW-complexes with fundamental group G into simple homotopy types is related to the size of matrices over ZG needed to represent all elements of the Whitehead group of G (see [9, Theorem 41). For other purposes O.S. Rothaus (in [18]) has displayed units representing the Whitehead group of dihedral groups of order 2p, for odd primes p. In this paper, Rothaus’ result is generalized and applied to simple homotopy classification. We begin in the next section with a general discussion of the representation of KIZG by units of ZG. The two sections which follow describe alternative norms and their connection with the reduced norm. The units representing KlZG for certain metacyclic groups G are then given. The last section is devoted to topological applications. The following notation, conventions and definitions are used. Rings R are associative with 1. The center of R is T(R). The group of units in R is GL1(R) = R*. Elementary n x n matrices are those in the subgroup E,(R) of GL, (R) generated by matrices which equal the identity matrix except possibly at a single off-diagonal entry. The stabilization maps: M,(R) --, M,,+,(R) take a matrix M to

Local Area Networks: Software and Related Issues

In this paper, we present a review of the issues that affect the software requirements for a local area network. We introduce protocols for the local area networks and characterize their software needs. Two approaches to operating systems are outlined and examples of each approach are presented. Various applications which use local area networks and performance issues are also discussed.

Mutual consistency in decentralized distributed systems

In this paper we set forth a simple and efficient algorithm for managing replicated data in a decentralized distributed system, which allows for inserts, deletes, updates, and synonyms and which achieves a high degree of availability in the face of node or communication failures. We focus on the approach developed recently by Fischer and Michael and exploit the knowledge of the semantics of the database operations.

A Scheme of Parallel Processing for MIMD Systems

This paper presents a recognition procedure for parallel tasks in the user program written in a conventional programming language. To establish our program model, it describes the parallelism of the program in tenns of a process flow graph in which the relationships among processes are of predecessors and successors. And finally it presents a parallel processing scheme which realizes automatically the recognition of parallel tasks and schedules these tasks for parallel execution.

Universal and Representative Instances Using Unmarked Nulls

Representative instances are important since in addition to the universal relation view they permit updates in the individual relations; if these updates are locally valid, then they do not violate global consistency. In this paper, we first give conditions under which we can construct a universal instance I which uses only one type of nulls (unmarked nulls) and whose total projections onto the relation schemes yield exactly the database relations. Then, we give a necessary and sufficient condition under which this I is actually a representative instance. Our conditions are simple and easily maintained.

Introduction to the special issue on the use of entity-relationship concepts in databases and related software

This is a collection of selected papers from the Third International Conference on Entity-Relationship (ER) Approach. These papers deal with the development of both principles and pragmatics of the ER approach in computer data engineering and reflect the trends of the expanding and maturing nature of the ER concept in the areas of database logical design, data management and control, database-oriented tools development, and reliability control in distributed database systems. The first five papers address the theory of entity relationships and foundations of the field of the ER approach. In conceptual schema design of databases, both structural and behavioral analyses of data based on the ER approach are discussed separately by Kent, Jajodia and Ng, and Sakai. The paper by Kent describes a methodology for data analysis leading to logical database design, based upon identifying the facts about entities that are to be maintained in a database, in terms of relationships among entities that are aggregated into data records: one record per fact rather than one record per entity. In contrast, most methodologies focus on entities and suggest that data represents two kinds of facts: facts about things (attributes) and facts that connect things (relationships). This implies that the distinction between relationships and attributes has to be made in the design process and design schemes have to proceed in two stages, one for handling relationships and the other for handling attributes. Above all, most methodologies assume that a simple representation is available for all entities. However, in reality, represen-