Interconnection Networks

One of the greatest challenges faced by designers of digital systems is optimizing the communication and interconnection between system components. Interconnection networks offer an attractive and economical solution to this communication crisis and are fast becoming pervasive in digital systems. Current trends suggest that this communication bottleneck will be even more problematic when designing future generations of machines. Consequently, the anatomy of an interconnection network router and science of interconnection network design will only grow in importance in the coming years.This book offers a detailed and comprehensive presentation of the basic principles of interconnection network design, clearly illustrating them with numerous examples, chapter exercises, and case studies. It incorporates hardware-level descriptions of concepts, allowing a designer to see all the steps of the process from abstract design to concrete implementation. Case studies throughout the book draw on extensive author experience in designing interconnection networks over a period of more than twenty years, providing real world examples of what works, and what doesn't. Tightly couples concepts with implementation costs to facilitate a deeper understanding of the tradeoffs in the design of a practical network. A set of examples and exercises in every chapter help the reader to fully understand all the implications of every design decision.

Foreword -- Foreword to the First Printing -- Preface -- Chapter 1 -- Introduction -- Chapter 2 -- Message Switching Layer -- Chapter 3 -- Deadlock, Livelock, and Starvation -- Chapter 4 -- Routing Algorithms -- Chapter 5 -- CollectiveCommunicationSupport -- Chapter 6 -- Fault-Tolerant Routing -- Chapter 7 -- Network Architectures -- Chapter 8 -- Messaging Layer Software -- Chapter 9 -- Performance Evaluation -- Appendix A -- Formal Definitions for Deadlock Avoidance -- Appendix B -- Acronyms -- References -- Index.

Programmable Logic Devices (PLDs) have become the key implementation medium for the vast majority of digital circuits designed today. While the highest-volume devices are still built with full-fabrication rather than field programmability, the trend towards ever fewer ASICs and more FPGAs is clear. This makes the field of PLD architecture ever more important, as there is stronger demand for faster, smaller, cheaper and lower-power programmable logic. PLDs are 90% routing and 10% logic. This book focuses on that 90% that is the programmable routing: the manner in which the programmable wires are connected and the circuit design of the programmable switches themselves. Anyone seeking to understand the design of an FPGA needs to become lit erate in the complexities of programmable routing architecture. This book builds on the state-of-the-art of programmable interconnect by providing new methods of investigating and measuring interconnect structures, as well as new programmable switch basic circuits. The early portion of this book provides an excellent survey of interconnec tion structures and circuits as they exist today. Lemieux and Lewis then provide a new way to design sparse crossbars as they are used in PLDs, and show that the method works with an empirical validation. This is one of a few routing architecture works that employ analytical methods to deal with the routing archi tecture design. The analysis permits interesting insights not typically possible with the standard empirical approach.

The advent of very large scale integrated circuit technology has enabled the construction of very complex and large interconnection networks. By most accounts, the next generation of supercomputers will achieve its gains by increasing the number of processing elements, rather than by using faster processors. The most difficult technical problem in constructing a supercom puter will be the design of the interconnection network through which the processors communicate. Selecting an appropriate and adequate topological structure of interconnection networks will become a critical issue, on which many research efforts have been made over the past decade. The book is aimed to attract the readers' attention to such an important research area. Graph theory is a fundamental and powerful mathematical tool for de signing and analyzing interconnection networks, since the topological struc ture of an interconnection network is a graph. This fact has been univer sally accepted by computer scientists and engineers. This book provides the most basic problems, concepts and well-established results on the topological structure and analysis of interconnection networks in the language of graph theory. The material originates from a vast amount of literature, but the theory presented is developed carefully and skillfully. The treatment is gen erally self-contained, and most stated results are proved. No exercises are explicitly exhibited, but there are some stated results whose proofs are left to the reader to consolidate his understanding of the material.

This volume provides an introduction to and overview of the emerging field of interconnected networks which include multilayer or multiplex networks, as well as networks of networks. Such networks present structural and dynamical features quite different from those observed in isolated networks. The presence of links between different networks or layers of a network typically alters the way such interconnected networks behave – understanding the role of interconnecting links is therefore a crucial step towards a more accurate description of real-world systems. While examples of such dissimilar properties are becoming more abundant – for example regarding diffusion, robustness and competition – the root of such differences remains to be elucidated. Each chapter in this topical collection is self-contained and can be read on its own, thus making it also suitable as reference for experienced researchers wishing to focus on a particular topic.

This unique text/reference provides an overview of crossbar-based interconnection networks, offering novel perspectives on these important components of high-performance, parallel-processor systems. A particular focus is placed on solutions to the blocking and scalability problems. Topics and features: introduces the fundamental concepts in interconnection networks in multi-processor systems, including issues of blocking, scalability, and crossbar networks; presents a classification of interconnection networks, and provides information on recognizing each of the networks; examines the challenges of blocking and scalability, and analyzes the different solutions that have been proposed; reviews a variety of different approaches to improve fault tolerance in multistage interconnection networks; discusses the scalable crossbar network, which is a non-blocking interconnection network that uses small-sized crossbar switches as switching elements. This invaluable work will be of great benefit to students, researchers and practitioners interested in computer networks, parallel processing and reliability engineering. The text is also essential reading for course modules on interconnection network design and reliability.

The advancement of large scale integrated circuit technology has enabled the construction of complex interconnection networks. Graph theory provides a fundamental tool for designing and analyzing such networks. Graph Theory and Interconnection Networks provides a thorough understanding of these interrelated topics. After a brief introduction to graph terminology, the book presents well-known interconnection networks as examples of graphs, followed by in-depth coverage of Hamiltonian graphs. Different types of problems illustrate the wide range of available methods for solving such problems. The text also explores recent progress on the diagnosability of graphs under various models.

This book introduces different interconnection networks applied to different systems. Interconnection networks are used to communicate processing units in a multi-processor system, routers in communication networks, and servers in data centers. Queuing techniques are applied to interconnection networks to support a higher utilization of resources. There are different queuing strategies, and these determine not only the performance of the interconnection network, but also the set of requirements to make them work effectively and their cost. Routing algorithms are used to find routes to destinations and directions in what information travels. Additional properties, such as avoiding deadlocks and congestion, are sought. Effective routing algorithms need to be paired up with these networks. The book will introduce the most relevant interconnection networks, queuing strategies, and routing algorithm. It discusses their properties and how these leverage the performance of the whole interconnection system. In addition, the book covers additional topics for memory management and congestion avoidance, used to extract higher performance from the interconnection network.