Michel Raynal Libros

Distributed computing is essential for applications that involve entities like processes, peers, or agents, each with limited knowledge of input parameters. Unlike parallel computing, which focuses on efficiency, and real-time computing, which emphasizes timely execution, distributed computing addresses the uncertainty stemming from multiple control flows, asynchronous communication, and dynamic behaviors. Although some distributed algorithms may appear simple, their behavior can be complex and challenging to analyze. This book aims to comprehensively present the fundamental concepts and algorithms of distributed computing, focusing on entities that communicate via asynchronous networks. It is organized into seventeen chapters across six parts: distributed graph algorithms, logical time and global states, mutual exclusion and resource allocation, high-level communication abstractions, distributed property detection, and distributed shared memory. Each chapter has clear objectives and is enriched with examples, summaries, exercises, and annotated bibliographies. This work serves as an introduction to distributed computing, targeting advanced undergraduates, graduate students in computer science and mathematics, and professionals involved in designing distributed applications. A basic understanding of algorithms and operating systems is recommended for readers.

Focusing on fault-tolerant distributed programming, the book introduces essential abstractions and algorithms crucial for reliable communication and consensus in distributed applications. It addresses challenges like asynchrony, process crashes, message losses, and Byzantine faults. The author presents concepts incrementally, beginning with clear specifications and progressing to intuitive algorithm descriptions, culminating in formal correctness proofs. This structured approach equips software designers and programmers with the tools needed to navigate complex distributed systems effectively.

This book explains synchronization and the implementation of concurrent objects, presenting synchronization algorithms while also introducing the theory that underlies the implementation of concurrent objects in the presence of asynchrony and process crashes.

Focusing on the foundational concepts of computing, this monograph explores the theoretical tools essential for understanding system design amidst technological evolution. It delves into asynchronous read/write shared memory systems, addressing challenges like process crashes and anonymity. Key topics include universal constructions, consensus numbers, and distributed recursivity. The book presents various distributed algorithms, enhancing comprehension of these concepts while showcasing their elegance and complexity. It aims to equip readers with a solid grasp of what is achievable within different computing models.

Focusing on the complexities of distributed computing, this book addresses the uncertainties of asynchrony and process failures in message-passing systems. It introduces essential communication abstractions, such as the register and reliable broadcast abstractions, which facilitate consistent communication among processes. The text also explores consensus agreement abstractions that enable cooperation in the face of failures. By emphasizing failure detectors for fault tolerance, it provides a framework for designing distributed software with provable properties, while discussing associated impossibility results.

Focusing on the challenges of uncertainty in synchronous message-passing systems, this book delves into the complexities of process failures, including crash, omission, and Byzantine behaviors. It addresses fundamental agreement problems essential for distributed systems, such as consensus, interactive consistency, and non-blocking atomic commit. By presenting algorithms and computability bounds for these issues, the book equips application designers with the tools to ensure cooperation and agreement despite failures, enabling the development of reliable distributed software.