Browse Prior Art Database

TRANSACTIONS AND CONSISTENCY IN DISTRIBUTED DATABASE SYSTEMS

IP.com Disclosure Number: IPCOM000148821D
Original Publication Date: 1979-Jun-05
Included in the Prior Art Database: 2007-Mar-30
Document File: 20 page(s) / 1M

Publishing Venue

Software Patent Institute

Related People

Traiger, Irving L.: AUTHOR [+5]

Abstract

Irving L. Traiger

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 7% of the total text.

Page 1 of 20

~ 5 2 5 5 5

     (33155) 6/5/79
Computer Science

TRANSACTIONS AND CONSISTENCY IN DISTRIBUTED DATABASE SYSTEMS

Irving L. Traiger
James N. Gray
Cesare A. Galtieri
Bruce G. Lindsay

IBM Research Laboratory
San Jose, California 95193

ABSTRACT: The notions of transaction and of data consistency are defined
for a distributed system. The cases of partitioned data, where a given
file is stored only at one node, and replicated data, where the contents
of a file may be replicated at several nodes are discussed It is argued
that the distribution and replication of data should be transparent to
the programs which use the data. Techniques for providing such trans-
parency are abstracted and discussed.

By extending the notions of system schedule and system clock to handle
multiple nodes, it is shown that consistency for partitioned data is
amenable to the techniques used for centralized data.

Research Division

IBM Yorktown Heights, New York . San Jose, California Zurich, Switzerland

[This page contains 1 picture or other non-text object]

Page 2 of 20

[This page contains 1 picture or other non-text object]

Page 3 of 20

TRANSACTIONS AND CONSISTENCY IN DISTRIBUTED DATABASE SYSTEMS

Irving L. Traiger, James N. Gray, Cesare A. Galtieri, Bruce G. Lindsay

IBM Research Division San Jose Laboratory San Jose, California

ABSTRACT: The notions of transaction and of data consistency are defined for a distributed system. The cases of partitioned data, where a given file is stored only at one node, and replicated data, where the contents of a file may be replicated at several nodes are discussed. It is argued that the distribution and replication of data should be transparent to the programs which use the data. Techniques for providing such transparency are abstracted and discussed.

By extending the notions of system schedule and system clock to handle multiple nodes, it is shown that consistency for partitioned data is amenable to the techniques used for centralized data.

[This page contains 1 picture or other non-text object]

Page 4 of 20

INTRODUCTION

To our knowledge, no general purpose distributed system provides the notion of ' network job ' , a co-ordinated unit of work which operates at several nodes. Rather, a task which operates at several nodes must be carefully programmed to be sensitive to data location in the network and to node and network failures. It is our thesis that the difficulty of constructing such programs is a principal cause for the dearth of systems which do distributed processing.

We conjecture that the notion of transaction as used in most data management systems general- izes to the network environment. This paper suggests that network systems should provide the notion of a transaction as an abstraction which eases the construction of programs in a distribut- ed system. Transactions would provide the programmer with the following...