Network Algorithmics: An Interdisciplinary Approach to Designing Fast Networked Devices

ISBN-13: 9780120884773 / Angielski / Twarda / 2004 / 496 str.

George Varghese

Opis

In designing a network device, you make dozens of decisions that affect the speed with which it will perform-sometimes for better, but sometimes for worse. Network Algorithmics provides a complete, coherent methodology for maximizing speed while meeting your other design goals. Author George Varghese begins by laying out the implementation bottlenecks that are most often encountered at four disparate levels of implementation: protocol, OS, hardware, and architecture. He then derives 15 solid principles-ranging from the commonly recognized to the groundbreaking-that are key to breaking these bottlenecks. The rest of the book is devoted to a systematic application of these principles to bottlenecks found specifically in endnodes, interconnect devices, and specialty functions such as security and measurement that can be located anywhere along the network. This immensely practical, clearly presented information will benefit anyone involved with network implementation, as well as students who have made this work their goal. FOR INSTRUCTORS: To obtain access to the solutions manual for this title simply register on our textbook website (textbooks.elsevier.com)and request access to the Computer Science subject area. Once approved (usually within one business day) you will be able to access all of the instructor-only materials through the "Instructor Manual" link on this book's academic web page at textbooks.elsevier.com.

• Addresses the bottlenecks found in all kinds of network devices, (data copying, control transfer, demultiplexing, timers, and more) and offers ways to break them
• Presents techniques suitable specifically for endnodes, including Web servers
• Presents techniques suitable specifically for interconnect devices, including routers, bridges, and gateways
• Written as a practical guide for implementers but full of valuable insights for students, teachers, and researchers
• Includes end-of-chapter summaries and exercises

Recenzje

"George Varghese has had a remarkable impact on the real world of networking with his algorithmic innovations over many years. The networking research and development community is fortunate that he has now distilled his knowledge in this very readable, insightful, and much-needed book." --Bruce Davie, Cisco Fellow, Cisco Systems

"This book nicely describes implementation tricks for building fast networking stacks, particularly in routers. This is a much needed book, I don't know of any other that covers this sort of implementation advice. George Varghese has invented several techniques to help speed up the Internet and in his book he provides interesting insight into this, and much more." --Radia Perlman, Distinguished Engineer, Sun Microsystems

Spis treści

1 Introducing Network Algorithmics
2 Network Implementation Models
3 Fifteen Implementation Principles
4 Principles in Action
5 Copying Data
6 Transfering Control
7 Maintaining Timers
8 Demultiplexing
9 Protocol Processing
10 Exact Match Lookups
11 Prefix Match Lookups
12 Packet Classification
13 Switching
14 Scheduling Packets
15 Routers as Distributed Systems
16 Measuring Network Traffic
17 Network Security
18 Conclusions
Appendix A Detailed Models

O autorze

George Varghese is a widely recognized authority on the art of network protocol implementation. Currently a Principal Researcher and Partner at Microsoft Research, he has served as a professor in the Department of Computer Science at UC-San Diego, he has previously taught at Washington University and Stanford University. Elected a Fellow of the ACM in 2002, he holds (with colleagues) 22 patents in the general field of network algorithmics. Several algorithms that he helped develop have found their way into commercial systems, including Linux (timing wheels), the Cisco GSR (DRR), and MS Windows (IP lookups). Varghese has written more than 100 papers on networking, computer architecture, genomics, and databases.