802.1aq Shortest Path Bridging Design and Evolution E-Book

David Allan

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 334

Veröffentlichungsjahr: 2012

Beschreibung

Facilitates both the understanding and adoption of 802.1aq as a networking solution

802.1aq Shortest Path Bridging (SPB) is a technology that greatly simplifies the creation and configuration of carrier, enterprise, and cloud computing networks—by using modern computing power to deprecate signaling, and to integrate multicast, multipath routing, and large-scale virtualization. It is arguably one of the most significant enhancements in Ethernet's history.

802.1aq Shortest Path Bridging Design and Evolution explains both the "what" and the "why" of the technology standard being set today. It covers which decisions were elective and which were dictated by the design goals by using a multipart approach that first explains what SPB is, before transitioning into narrative form to describe the design processes and decisions behind it.

To make SPB accessible to the data networking professional from multiple perspectives, the book:

• Provides a "Reader's Companion" to the standard

• Dissects the different elements of SPB

• Offers applications and potential futures for the technology

802.1aq Shortest Path Bridging Design and Evolution will appeal to system implementers, system and network architects, academics, IT professionals, and general networking professionals.

Leseprobe

Table of Contents

Cover

Title page

Copyright page

Figures

Acknowledgments

Introduction

Abbreviations

CHAPTER 1 IEEE 802.1aq in a Nutshell: Antecedents and Technology

SPB: ANTECEDENTS AND PRINCIPLES OF NETWORK OPERATION

SPB TECHNOLOGY: THE CONTROL PLANE

SPB TECHNOLOGY: PATH COMPUTATION

SPB TECHNOLOGY: LOOP AVOIDANCE

SUMMARY

CHAPTER 2 Why SPB Looks as It Does

THE PROBLEM SPACE

HISTORY

LYNCHPINS: CONSTRAINTS WE CHOSE TO RESPECT

REINTERPRETING WHAT ALREADY EXISTS IN ETHERNET

MEANING OF MAC ADDRESSES IN PBB-TE AND SPBM; PORT AND NODAL MACS

ROUNDING OUT DESIGN DETAILS

CHAPTER 3 Why the SPB Control Plane Looks as It Does

THE CONTROL PLANE IS AS SIMPLE AS IT CAN BE, BUT NO SIMPLER

CONTROL PLANE INFORMATION

CONTROL PLANE: ALGORITHM ASPECTS

THE CO NATURE OF ETHERNET AND ITS IMPLICATIONS FOR ROUTING

CHAPTER 4 Practical Deployment Considerations

IN-SERVICE UPGRADE AND SERVICE MIGRATION

DUAL HOMING

SHARED SEGMENTS

CHAPTER 5 Applications of SPB

SPB IMPLEMENTATION OF MEF SERVICES AND METRO INFRASTRUCTURE

THE DATA CENTER AND GENERAL ENTERPRISE APPLICATIONS

RADIO ACCESS NETWORKS (RANS)

MULTICARRIER CONSIDERATIONS

CHAPTER 6 Futures

FURTHER RESEARCH ON LOAD SPREADING ALGORITHMS

LAYER 3 INTEGRATION WITH SPBM

MULTIAREA

EXTENDED CONNECTIVITY MODELS: SPANNING TREES

EXTENDED CONNECTIVITY MODELS: NONPLANAR GRAPHS

Conclusion

References

Index

IEEE and 802 are registered trademarks of the Institute of Electrical and Electronics Engineers, Incorporated (www.ieee.org/).

IEEE Standards designations are trademarks of the Institute of Electrical and Electronics Engineers, Incorporated (www.ieee.org/).

Non-IEEE trademarks are the respective property of their owners.

Copyright © 2012 by the Institute of Electrical and Electronics Engineers. All rights reserved.

3 Park Avenue, New York, NY 10016-5997, USA

Published by John Wiley & Sons, Inc., Hoboken, New Jersey.

Published simultaneously in Canada.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, [phone], fax [phone], or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, [phone], fax [phone], or online at http://www.wiley.com/go/permissions.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at [phone], outside the United States at [phone] or fax [phone].

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

See other IEEE Standards and standards-related product listings at: http://standards.ieee.org.

Library of Congress Cataloging-in-Publication Data is available.

ISBN: 978-1-118-14866-2

Figures

Figure 1.1

A simple spanning tree

Figure 1.2

The evolution of Ethernet stacking

Figure 1.3

IEEE 802.1aq technology—data plane connectivity for BEB “A”

Figure 1.4

New information items in IS-IS for SPBM

Figure 2.1

Metro network architecture: layer view

Figure 2.2

Hop-by-hop load spreading—ECMP

Figure 2.3

End-to-end load spreading—ECMT

Figure 2.4

Port-based interface structures (from [PBB])

Figure 2.5

Different sources may not exhibit congruent trees                                

Figure 2.6

The condition for downstream congruency

Figure 3.1

The rise of computing power

Figure 3.2

Signaling to support multipoint-to-multipoint tree setup

Figure 3.3

Minimum-cost MDT

Figure 3.4

Shortest path tree with maximum diversity

Figure 3.5

Minimum-cost shortest path tree

Figure 3.6

A multicast reconvergence scenario

Figure 3.7

New information items for IS-IS for SPBM

Figure 3.8

New information items for IS-IS for SPBV

Figure 3.9

Use of multiple I-SIDs for resilient E-TREE structures

Figure 3.10

The routed restoration cycle

Figure 3.11

The importance of consistent tiebreaking

Figure 3.12

A looping example

Figure 3.13

The SPB topology digest

Figure 3.14

“Fat tree” switching structure

Figure 3.15

Network partitioning from the perspective of node 2

Figure 3.16

Network coalescence from the perspective of node 2

Figure 4.1

Announcement of a dual-homed UNI into SPBM

Figure 4.2

Announcement of a UNI fault into SPBM

Figure 4.3

SPBM delivering “multipoint link aggregation”

Figure 4.4

Multipoint LAG endpoints: structure and advertisements

Figure 4.5

Multipoint LAG: failure of an access link

Figure 4.6

IS-IS use of the pseudonode to model broadcast segments

Figure 4.7

The ease of loop formation on LAN segments

Figure 4.8

SPBM overlay of bridged Ethernet emulation

Figure 4.9

LAN segments and frame duplication

Figure 4.10

Use of multiple Ethernet segments by SPBM overlay

Figure 6.1

A multiarea model potentially used by SPB

Figure 6.2

Multiarea forwarding and the common (unary) FDB model

Figure 6.3

SPBM ABB structures

Figure 6.4

PCI: PBB-TE trunk between individual endpoints

Acknowledgments

This document is an amalgam of practical wisdom of Dave Allan, Peter Ashwood-Smith, Nigel Bragg, Janos Farkas, Don Fedyk, Jérôme Chiabaut, Dinesh Mohan, Mick Seaman, and Paul Unbehagen.

Thanks also to Anne Bragg for her persistence in reducing our wayward capitalization, punctuation, and other stylistic idiosyncracies to more conventional forms.

Technical ReviewersSimon Parry, Ciena LimitedJoel M. Halpern, Distinguished Engineer, Ericsson

Introduction

Ethernet is a difficult and demanding taskmaster.

We start from the position that for any networking technology of sufficient power, an elegant and self-consistent solution to a given connectivity problem exists. Switched Ethernet is the product of 30 years on Occam’s razor, and although the attributes and scale of the application domains covered by this book were until recently inconceivable, we have concluded that it remains a technology of sufficient power and self-consistency.

The success and longevity of Ethernet can be put down to the fact that it has been able to evolve to accommodate new requirements, both in its original LAN application space and in the increasing proportion of Provider networking space. Shortest path bridging (SPB) is one of the most recent of these evolutionary steps, and we would like to establish at this early point both what is the fundamental problem it solves and why the solution is useful.

The short and sufficient answer is, “elimination of the Spanning Tree Protocol and its shortcomings, and its replacement by a superior routed technology, and without changing the service model.” This answer is “sufficient” for now because it is generally accepted in the industry that Spanning Tree Protocol presents problems and limits the applications accessible to Ethernet, and we therefore defer further discussion on the origins and root cause of this problem to the beginning of the next section.

Replacement of Spanning Tree Protocol by something substantially superior is a general “good” that applies to Ethernet networking in both Enterprise and Provider space. The other key requirement of Ethernet networking, which is increasingly shared by Enterprise applications as well as Providers, is virtualization, which is the ability to support multiple independent LAN segments on the same physical infrastructure. SPB did not originate the technology to do this, but directly supports earlier IEEE Standards (Provider Bridging and Provider Backbone Bridging) that defined the hierarchical data path constructs to support virtualization.

There are two variants of SPB, one using the 802.1ad Q-in-Q datapath—shortest path bridging VID (SPBV)—and one using the hierarchical 802.1ah MAC-in-MAC datapath—shortest path bridging MAC (SPBM). SPBV and SPBM share a control plane, algorithms, and common routing mechanisms; where the term “SPB” is used, this describes aspects common to both variants.

The authors embarked on their journey motivated by the issues of highly scalable networks intended for deployment by Service Providers, a path that lead to a precursor of SPBM, known as Provider Link State Bridging, or PLSB. A significant number of the topics discussed are more relevant to a technology supporting virtualization, and the reader should therefore expect a significant focus on SPBM, both because of its highly scalable support of virtualization and, as a corollary, because of from where the authors came.

The first, and substantially shorter, part of this book summarizes succinctly and informally what SPB “is” today, with the aim of offering a reader new to the technology a consistent mental model of what it does and how it does it. The second part provides the rationale for why SPB is as it is, and has to be so. This therefore not only includes a post hoc rationalization of SPB with the 20:20 vision of hindsight, but also describes some of the blind alleys explored in getting there, because these alleys give additional insight into why SPB has to be what it is.

We start with a short history of SPB and its antecedents, with only the briefest allusions to the motivations for SPB at this stage (Chapter 1, “IEEE 802.1aq in a Nutshell; Antecedents and Technology”). We then offer a short description of SPB as it is now (three sections on “SPB technology: The Control Plane,” starting on p. 15). These are succinct, but capture the key principles and attributes of SPB. We nonetheless anticipate that readers will finish this with more questions than answers, such as:

Why

is

congruence so important?

How big a network can you really make?

Why is this really so different from other network technologies?

The rest of the book sets out to answer these questions.

We start by considering the key requirements that the different networking scenarios present (the section on “The Problem Space,” p. 37). As a way of introducing the constraints and degrees of freedom offered by the Ethernet baseline, we follow this with a summary of our progress toward SPB as documented here (“History,” p. 52). Unlike subsequent chapters which organize material by topic, this is a chronological record that talks through some of the twists on the journey.

We then revisit some fundamental principles of Ethernet, showing why we decided to stick with some key bridging constructs even though the adoption of a control plane meant that they were no longer strictly mandatory. We also reinterpret the use of some other key Ethernet concepts, such as the Virtual LAN (VLAN), always within their strict specifications, but in ways possibly contrary to received wisdom on their usage.

In the section on “Rounding Out Design Details” (p. 69), we focus on overall networking challenges beyond basic functionality that have to be addressed to make a technology deployable. There is a discussion of data plane instrumentation, the OAM (deliberately short, to make an important point), followed by more extensive discussion on dual-homing for resiliency, always a thorny issue for Ethernet.

The section in Chapter 3 on “The Control Plane Is as Simple as It Can Be, but No Simpler” (p. 74) shows that essentially all SPB functionality can be delivered by the routing system. We first discuss SPB’s most radical departure from previous received wisdom, the complete elimination of signaling from both unicast and multicast state installation. We then provide a factual introduction to the extensions to IS-IS required by SPB, showing how modest these are. Finally, we explain some of the algorithmic innovations required by SPB over previous link state routing practice.

So far, the exposition has assumed point-to-point connectivity between bridges in the SPB domain, and ignored the traditional shared segment. Chapter 4, “Practical Deployment Considerations” (p. 130) considers this and other topics, because an SPB overlay of an emulated LAN segment is a real deployment scenario. Although a solution is described, this is not quite a “done deal,” because it needs modest extensions to the Ethernet forwarding path.

Next, in Chapter 5, we explore applications of SPB (p. 150), providing walk-through examples of operation in various deployment scenarios covering the delivery of Metro Ethernet Forum defined services by carriers, and the use of SPB in enterprise applications. Because Ethernet has been very widely deployed in these applications in the past, this treatment focuses on the ability of SPM to address the limitations and deficiencies of earlier Ethernet technologies.

Finally, in Chapter 6, we explore whole new capabilities that SPBM could be extended to provide.

In SPBM, the service primitive is the emulated LAN segment. The LAN segment at Layer 2 is the IP Subnet at Layer 3, and IS-IS has been routing IP for many years. So, if IS-IS retains its IP personality as well as running SPBM, we have a single control plane with a complete view of both Layer 2 and Layer 3 topologies, and we can “route at the edge, switch through the core,” and virtualize the notion of location implied by the subnet prefix. The SPBM service architecture can now be used to construct a virtual network of IP Subnets, and the result is IP-VPN capability as well as the native virtual LAN segments. This is elaborated in the section on “Layer 3 Integration with SPBM.”

We also investigate how the “Multiarea” capability of IS-IS can be applied to SPB. Finally, we explore how the shortest path tree may be extended with other connectivity styles and incorporated into the framework supported by the control plane (the section on “Extended Connectivity Models: Spanning Trees”). We first show how traditional spanning trees may be constructed. We then turn our attention to the coercion of traffic off shortest paths, for traffic engineering purposes, without causing undesirable side effects within the routed system.

We started by asserting that the success of Ethernet and its evo­lution is the consequence of 30 years on Occam’s razor and that we have discovered that it is a “technology of sufficient power and self-onsistency.” We hope the reader in the process of the journey of discovery outlined above ultimately agrees with us.

As an early hint as to why we believe this, Ethernet descended from a broadcast medium. This is very important, as the types of connectivity offered by Ethernet are derived from filtering of the basic broadcast behavior, with the point-to-point connection simply being the most extreme form of filtering. The implication here is that all types of communication—one-to-all, one-to-some, and one-to-one—can be derived from the basic transmission behavior combined with filtering. This is distinctly different from the history of most other network technologies, which have started from one-to-one connections as the service primitive, and subsequently overlaid broadcast behavior onto this unicast model.

We also attach much importance to the fact that Ethernet uses global addressing in the data plane, a characteristic that it shares with only one other major production technology, IP. Everything seems to follow from this choice, rather than the adoption of a “link-local” identifier as a forwarding scheme:

with the proper control plane it scales. Scalability is in practice dominated by state volume, and not theoretical considerations of addressing space; IP uses address aggregation to control the issues of global addresses; Ethernet uses

hierarchy

, the alternative route to scale.

global data plane identifiers make a frame self-describing and remove the need for rafts of complexity; signaling can be eliminated because global identifier information can be communicated by more efficient means; whole classes of subtle errors caused by lack of synchronization between control and forwarding planes are eliminated; and the OAM to detect the remaining fault classes is much simpler.

Ethernet continues to evolve, and this book is simply a snapshot of a point on the journey. We have endeavored to provide insight into what we believe to be a significant evolutionary step in Ethernet technology. Ethernet’s longevity and its ability to evolve to address new requirements are an independent testimony to its fundamental “fitness for purpose.” However, as it has evolved, the limitations of spanning tree have become increasingly apparent, and finally become a real barrier to further extensions to its scope. With SPB, Ethernet has acquired the state of the art in distributed routing technology, which is now available to future evolutionary developments.

Abbreviations

ABB

Area Boundary Bridge

AESA

ATM End System Address

ARP

Address Resolution Protocol

ATM

Asynchronous Transfer Mode

BCB

Backbone Core Bridge

BEB

Backbone Edge Bridge

BGP

Border Gateway Protocol

BNG

Broadband Network Gateway

BRAS

Broadband Remote Access Server

CLIP

Classical IP over ATM

CSNP

Complete Sequence Number Packet

CTO

Chief Technology Office/Officer

DA

Destination MAC address

DSLAM

Digital Subscriber Loop Access Multiplexer

ECMP

Equal Cost Multi Path

ECT

Equal Cost Tree

EMS

Element Management System

ESP

Ethernet Switched Path

IETF

Internet Engineering Task Force

IGMP

Internet Group Management Protocol

IP

Internet Protocol

I-SID

I Component Service ID

IS-IS

Intermediate System to Intermediate System

ITU-T

International Telecommunications Union—Telecommunications Standardization Sector

IVL

Independent VLAN Learning

LAN

Local Area Network

LDP

Label Distribution Protocol

LSP

Link State Packet

LTM

Link trace message

MAC

Media Access Control

MEF

Metro Ethernet Forum

MEL

Maintenance Entity Level

MEP

Maintenance End Point

MIP

Maintenance Intermediate Point

mp2p

Multipoint to point

MPLS

Multiprotocol Label Switching

NHRP

Next Hop Resolution Protocol

NNI

Network to Network Interface

NSAP

Network Service Access Point

OAM

Operations, Administration and Maintenance

OLT

Optical Line Termination

OUI

Organizationally Unique Identifier

p2mp

Point to Multipoint

PBBN

Provider Backbone Bridged Network

PBT

Provider Backbone Transport

PLSB

Provider Link State Bridging

PSNP

Partial Sequence Number Packet

RPFC

Reverse Path Forwarding Check

RT

Route Target

SA

Source MAC address

SDH

Synchronous Data Hierarchy

SONET

Synchronous Optical Network

SPB

Shortest Path Bridging

SPBM

Shortest Path Bridging MAC Mode

SPBV

Shortest Path Bridging VID mode

STP

Spanning Tree Protocol

SVC

Switched Virtual Circuit

TRILL

Transparent Connection of Lots of Links

TTL

Time to live

UNI

User to Network Interface

VID

VLAN ID

VLAN

Virtual LAN

VPLS

Virtual Private LAN Service

VPN

Virtual Private Network

VSI

Virtual Switch Instance

CHAPTER 1

IEEE 802.1aq in a Nutshell: Antecedents and Technology