Atomic Transactions
In Concurrent and Distributed Systems
- 1st Edition - August 1, 1993
- Imprint: Morgan Kaufmann
- Authors: Nancy A. Lynch, Michael Merritt, William E. Weihl, Alan Fekete
- Language: English
- Hardback ISBN:9 7 8 - 1 - 5 5 8 6 0 - 1 0 4 - 8
- eBook ISBN:9 7 8 - 0 - 0 8 - 0 4 9 9 5 4 - 3
This book develops a theory for transactions that provides practical solutions for system developers, focusing on the interface between the user and the database that executes tr… Read more
Purchase options
Institutional subscription on ScienceDirect
Request a sales quoteThis book develops a theory for transactions that provides practical solutions for system developers, focusing on the interface between the user and the database that executes transactions. Atomic transactions are a useful abstraction for programming concurrent and distributed data processing systems. Presents many important algorithms which provide maximum concurrency for transaction processing without sacrificing data integrity. The authors include a well-developed data processing case study to help readers understand transaction processing algorithms more clearly. The book offers conceptual tools for the design of new algorithms, and for devising variations on the familiar algorithms presented in the discussions. Whether your background is in the development of practical systems or formal methods, this book will offer you a new way to view distributed systems.
Atomic Transactions: In Concurrent and Distributed Systems
by Nancy Lynch, Michael Merritt, William Weihl and Alan Fekete
by Nancy Lynch, Michael Merritt, William Weihl and Alan Fekete
- Foreward
- Preface
- 1. Introduction
- 1.1 Distributed Systems
- 1.2 Atomic Transactions for Distributed Systems
- 1.2.1 Atomic Databases
- 1.2.2 Transactions
- 1.2.3 Nested Transactions
- 1.3 Transaction-Processing Algorithms
- 1.3.1 Locking
- 1.3.2 Timestamps
- 1.3.3 Optimistic Algorithms
- 1.3.4 Replication
- 1.3.5 Recovery
- 1.4 Formal Models
- 1.4.1 Why a Formal Model?
- 1.4.2 An Operational Automation Model
- 1.4.3 Correct Systems Simulate Serial System
- 1.5 Comparisons with the Classical Theory
- 1.5.1 Serializability
- 1.5.2 Classical Treatment of Recovery
- 1.5.3 Data Types and Nesting
- 1.5.4 Operational versus Axiomatic Model
- 1.5.5 A Finer Granularity of Analysis
- 1.6 Contents of This Book
- 1.6.1 The Basic Theory
- 1.6.2 Fundamental Transaction-Processing Algorithms
- 1.6.3 More Advanced Transaction-Processing Algorithms
- 1.7 Bibliographic Notes
- 1.7.1 Transactions
- 1.7.2 Correctness Conditions
- 1.7.3 Algorithms for Concurrency Control and Recovery
- 2 An Automation Model
- 2.1 Introduction
- 2.2 Action Signatures
- 2.3 Automata
- 2.3.1 Conventions: Transition Relations
- 2.3.2 More Examples of Automata
- 2.4 Executions and Behaviors
- 2.5 Composition
- 2.5.1 Compositiona of Action Signatures
- 2.5.2 Composition of Automata
- 2.5.3 Properties of Systems of Automata
- 2.6 Proofs about Automata
- 2.7 Relationships between Automata
- 2.7.1 Implementation
- 2.7.2 Possibilities Mappings
- 2.8 Preserving Properties
- 2.9 Discussion
- 2.9.1 Process Algebras
- 2.9.2 State-based Models
- 2.9.3 Uses of Models
- 2.10 Bibliographic Notes
- 2.11 Exercises
- 3 Serial Systems and Correctness
- 3.1 Introduction
- 3.2 System Types
- 3.3 General Structure of Serial Systems
- 3.4 Serial Actions and Well-Formedness
- 3.4.1 Basic Definitions
- 3.4.2 Well-Formedness
- 3.5 Serial Systems
- 3.5.1 Transaction Automata
- 3.5.2 Serial Object Automata
- 3.5.3 The Serial Scheduler
- 3.5.4 Serial Systems, Executions, Schedules and Behaviors
- 3.5.5 Convention: Fixing the System Type and Serial System
- 3.6 Atomicity
- 3.6.1 Atomicity of Action Sequences
- 3.6.2 Atomicity of Systems
- 3.6.3 Stronger Correctness Conditions
- 3.6.4 Atomicity and Intuition
- 3.6.5 Implications for Schedules and Executions
- 3.6.6 Correctness of Infinite Behaviors
- 3.7 An Additional Example
- 3.8 Bibliographic Notes
- 3.9 Exercises
- 4 Special Classes of Serial Systems
- 4.1 Introduction
- 4.2 Equieffectiveness
- 4.3 Commutativity
- 4.3.1 Forward Commutativity
- 4.3.2 Backward Commutativity
- 4.4 Special Classes of Operations
- 4.4.1 Transparent Operations
- 4.4.2 Obliterating Operations
- 4.5 Serial Dependency Relations
- 4.6 Bibliographic Notes
- 4.7 Exercises
- 5 The Atomicity Theorem
- 5.1 Introduction
- 5.2 Visibility
- 5.3 Simle Systems
- 5.3.1 Simple Database
- 5.3.2 Simple Systems, Executions, Schedules, and Behaviors
- 5.4 Event and Transaction Orders
- 5.4.1 SD-Affects Order
- 5.4.2 Affects Order
- 5.4.3 Sibling Orders
- 5.5 Atomicity Theorem and Proof Sketch
- 5.6 Proof of the Atomicity
- 5.6.1 Pictures
- 5.6.2 Behavior of Transactons
- 5.6.3 Behavior of Serial Objects
- 5.6.4 Behavior of the Serial Scheduler
- 5.6.5 Proof of the Main Result
- 5.7 Discussion
- 5.8 Bibliographic Notes
- 5.9 Exercises
- 6 Locking Algorithms
- 6.1 Introduction
- 6.2 Dyamic Atomicity
- 6.2.1 Completion Order
- 6.2.2. Generic Systems
- 6.2.3 Dyamic Atomicity
- 6.2.4 Local Dyamic Atomicity
- 6.3 General Commutativity-based Locking
- 6.3.1 Locking Objects
- 6.3.2 Correctness Proof
- 6.4 Moss's Algorithm
- 6.4.1 Moss Objects
- 6.4.2 Correctness Proof
- 6.4.3 Read/Write Locking Algorithm
- 6.5 General Undo Logging Algorithm
- 6.5.1 General Undo Logging Objects
- 6.5.2 Correctness Proof
- 6.6 Discussion
- 6.7 Bibliographic Notes
- 6.8 Exercises
- 7 Timestamp Algorithms
- 7.1 Introduction
- 7.2 Simple-Pseudotime Systems
- 7.2.1 Simple-Pseudotime Database
- 7.2.2 Simple-Pseudotime Systems, Executions, Schedules, and Behaviors
- 7.3 Static Atomicity
- 7.3.1 Pseudotime Order
- 7.3.2 Pseudotime Systems
- 7.3.3 Static Atomicity
- 7.3.4 Local Static Atomicity
- 7.4 Reed's Algorithm
- 7.4.1 Reed Objects
- 7.4.2 Correctness Proof
- 7.4.3 A Concrete Implementation
- 7.4.4 Correctness Proof
- 7.5 Type-specific Concurrency Control
- 7.5.1 Herlihy Objects
- 7.5.2 Correctness Proof
- 7.6 Discussion
- 7.7 Bibliographic Notes
- 7.8 Exercises
- 8 Hybrid Algorithms
- 8.1 Introduction
- 8.2 Hybrid Atomicity
- 8.2.1 Hybrid Systems
- 8.2.2 Hybrid Atomicity
- 8.2.3 Local Hybrid Atomicity
- 8.3 Dependecy-Based Hybrid Locking
- 8.3.1 Dependency Objects
- 8.3.2 Correctness Proof
- 8.4 Discussion
- 8.5 Bibliographic Notes
- 8.6 Exercises
- 9 Relationship to the Classical Theory
- 9.1 Introduction
- 9.2 Assumptions
- 9.2.1 Read/Write Serial Objects
- 9.2.2 Appropriate Return Values
- 9.2.3 A Sufficient Condition for Appropriate Return Values
- 9.3 The Serialization Graph Construction
- 9.4 Moss's Algorithm
- 9.5 Extension to General Data Types
- 9.5.1 Serialization Graphs
- 9.5.2 An Undo Logging Algorithm
- 9.6 Discussion
- 9.7 Bibliographic Notes
- 9.8 Exercises
- 10 Optimistic Algorithms
- 10.1 Introduction
- 10.2 Optimistic Hybrid Systems
- 10.2.1 Simple-Timestamp Systems
- 10.2.2 Optimistic Hybrid Systems
- 10.2.3 Optimistic Hybrid Atomicity
- 10.2.4 Local Optimistic Hybrid Atomicity
- 10.3 An Optimistic Dependency-based Algorithm
- 10.3.1 Optimistic Dependency Objects
- 10.3.2 Correctness Proof
- 10.4 Optimistic Pseudotime Systems
- 10.4.1 Optimisitic Pseudotimes Systems
- 10.4.2 Optimistic Static Atomicity
- 10.4.3 Local Optimistic Static Atomicity
- 10.5 An Optimistic Version fo Reed's Algorithm
- 10.5.1 Correctness Proof
- 10.6 Discussion
- 10.7 Bibliographic Notes
- 10.8 Exercises
- 11 Orphan Management Algorithms
- 11.1 Introduction
- 11.2 The Orphan Management Problem
- 11.3 An Affects Relation
- 11.4 Filtered Systems
- 11.4.1 The Filtered Database
- 11.4.2 The Filtered System
- 11.4.3 Simulation of the Generic System by the Filtered System
- 11.5 Piggyback Systems
- 11.5.1 The Piggyback Database
- 11.5.2 The Piggyback System
- 11.5.3 Simulation of the Generic System by the Piggyback System
- 11.6 Strictly Filtered Systems
- 11.6.1 The Strictly Filtered Database
- 11.6.2 The Strictly Filtered System
- 11.6.3 Simulation of the Generic System by the Clock System
- 11.7 Clock Systems
- 11.7.1 The Clock Database
- 11.7.2 The Clock System
- 11.7.3 Simulation of the Generic System by the Clock System
- 11.8 Discussion
- 11.9 Bibliographic Notes
- 11.10 Exercises
- 12 Replication
- 12.1 Introduction
- 12.2 Configurations
- 12.3 Fixed Configuration Quorum Consensus
- 12.3.1 System A, the Unreplicated Serial Sytem
- 12.3.2 System B, the Fixed Configuation Quorum Consensus Algorithm
- 12.3.3 Correctness Proof
- 12.4 Reconfigurable Quorum Consensus
- 12.4.1 System A, the Unreplicated Serial System with Dummy Reconfiguration
- 12.4.2 System B, the Reconfigurable Quorum Consensus Algorithm with a Centralized Configuration
- 12.4.3 Correctness Proof
- 12.4.4 System C, the Reconfiturable Quorum Consensus Algorithm with Replicated Configurations
- 12.4.5 Correctness Proof
- 12.5 Concurrent Replicated Systems
- 12.6 Discussion
- 12.7 Bibliographic Notes
- 12.8 Exercises
- A Mathematical Concepts
- A.1 Intorduction
- A.2 Sets
- A.3 Tuples
- A.4 Relations
- A.5 Functions and Partial Functions
- A.6 Trees
- A.7 Sequences
- Bibliography
- Index
- Edition: 1
- Published: August 1, 1993
- Imprint: Morgan Kaufmann
- Language: English
- Hardback ISBN: 9781558601048
- eBook ISBN: 9780080499543
NL
Nancy A. Lynch
About the author:Nancy A. Lynch is a professor of electrical engineering and computer science at MIT and heads MIT's Theory of Distributed Systems research group. She is the author of numerous research articles about distributed algorithms and impossibility results, and about formal modeling and verification of distributed systems.
MM
Michael Merritt
Michael Merritt, AT&T
WW
William E. Weihl
William E. Weihl, Massachusetts Institute of Technololgy
AF
Alan Fekete
Alan Fekete, University of Sydney