Storage Networks Explained E-Book

Ulf Troppens

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch
• Veröffentlichungsjahr: 2011

Seitenzahl: 901

Beschreibung

All you need to know about Storage Area Networks The amount of data of an average company doubles every year. Thus, companies who own 1TB of data today will own 32TB in five years. Storage networks help to tame such data quantities and to manage this data growth efficiently. Since stored data and information are the biggest asset of any company, anyone who is involved in the planning or the operation of IT systems requires a basic knowledge of the principle and the use of storage networks. Storage Networks Explained covers the fundaments, techniques and functions of storage networks such as disk subsystems, Fibre Channel SAN, Internet SCSI (iSCSI), Fibre Channel over Ethernet (FCoE), Network Attached Storage (NAS), file systems, and storage virtualization. Furthermore the authors describe the use of these techniques and how they are designed to achieve high-availability, flexibility, and scalability of data and applications. Additional attention is given to network backup and the management of storage networks. Written by leading experts in the field, this book on storage area networks is updated and fully revised. Key features: * Presents the basic concepts of storage networks, such as I/O techniques, disk subsystems, virtualization, NAS and SAN file systems * Covers the design of storage networks which provide flexible, highly-available, and scaleable IT systems * Explains the use of storage networks for data sharing, data protection, and digital archiving * Discusses management of storage networks using SNMP, SMI-S, and IEEE 1244 This book provides system administrators and system architects, as well as students and decision makers, with the tools needed for optimal selection and cost-effective use of storage networks. The Linux Journal awarded the first edition with the "Editor's Choice Award 2005" in the category "System Administration Book."

Leseprobe

Contents

About the Authors

Foreword to the Second Edition by Hermann Strass

Preface by the authors

WHAT DOES THIS BOOK DEAL WITH?

WHO SHOULD READ THIS BOOK?

HOW SHOULD THIS BOOK BE READ?

WHO HAS WRITTEN THIS BOOK?

ACKNOWLEDGEMENTS FOR THE SECOND EDITION

ACKNOWLEDGEMENTS FOR THE FIRST EDITION

List of Figures and Tables

FIGURES

TABLES

1 Introduction

1.1 SERVER-CENTRIC IT ARCHITECTURE AND ITS LIMITATIONS

1.2 STORAGE-CENTRIC IT ARCHITECTURE AND ITS ADVANTAGES

1.3 CASE STUDY: REPLACING A SERVER WITH STORAGE NETWORKS

1.4 THE STRUCTURE OF THE BOOK

Part I Technologies for Storage Networks

2 Intelligent Disk Subsystems

2.1 ARCHITECTURE OF INTELLIGENT DISK SUBSYSTEMS

2.2 HARD DISKS AND INTERNAL I/O CHANNELS

2.3 JBOD: JUST A BUNCH OF DISKS

2.4 STORAGE VIRTUALISATION USING RAID

2.5 DIFFERENT RAID LEVELS IN DETAIL

2.6 CACHING: ACCELERATION OF HARD DISK ACCESS

2.7 INTELLIGENT DISK SUBSYSTEMS

2.8 AVAILABILITY OF DISK SUBSYSTEMS

2.9 SUMMARY

3 I/O Techniques

3.1 THE PHYSICAL I/O PATH FROM THE CPU TO THE STORAGE SYSTEM

3.2 SCSI

3.3 THE FIBRE CHANNEL PROTOCOL STACK

3.4 FIBRE CHANNEL SAN

3.5 IP STORAGE

3.6 INFINIBAND-BASED STORAGE NETWORKS

3.7 FIBRE CHANNEL OVER ETHERNET (FCoE)

3.8 SUMMARY

4 File Systems and Network Attached Storage (NAS)

4.1 LOCAL FILE SYSTEMS

4.2 NETWORK FILE SYSTEMS AND FILE SERVERS

4.3 SHARED DISK FILE SYSTEMS

4.4 COMPARISON: FIBRE CHANNEL SAN, FCoE SAN, iSCSI SAN AND NAS

4.5 SUMMARY

5 Storage Virtualisation

5.1 ONCE AGAIN: VIRTUALISATION IN THE I/O PATH

5.2 LIMITATIONS AND REQUIREMENTS

5.3 DEFINITION OF STORAGE VIRTUALISATION

5.4 IMPLEMENTATION CONSIDERATIONS

5.5 STORAGE VIRTUALISATION ON BLOCK OR FILE LEVEL

5.6 STORAGE VIRTUALISATION ON VARIOUS LEVELS OF THE STORAGE NETWORK

5.7 SYMMETRIC AND ASYMMETRIC STORAGE VIRTUALISATION IN THE NETWORK

5.8 SUMMARY

Part II Application and Management of Storage Networks

6 Application of Storage Networks

6.1 DEFINITION OF THE TERM ‘STORAGE NETWORK’

6.2 STORAGE SHARING

6.3 AVAILABILITY OF DATA

6.4 ADAPTABILITY AND SCALABILITY OF IT SYSTEMS

6.5 SUMMARY

7 Network Backup

7.1 GENERAL CONDITIONS FOR BACKUP

7.2 NETWORK BACKUP SERVICES

7.3 COMPONENTS OF BACKUP SERVERS

7.4 BACKUP CLIENTS

7.5 PERFORMANCE GAINS AS A RESULT OF NETWORK BACKUP

7.6 PERFORMANCE BOTTLENECKS OF NETWORK BACKUP

7.7 LIMITED OPPORTUNITIES FOR INCREASING PERFORMANCE

7.8 NEXT GENERATION BACKUP

7.9 BACKUP OF FILE SYSTEMS

7.10 BACKUP OF DATABASES

7.11 ORGANISATIONAL ASPECTS OF BACKUP

7.12 SUMMARY

8 Archiving

8.1 TERMINOLOGY

8.2 MOTIVATION, CONDITIONS AND REQUIREMENTS

8.3 IMPLEMENTATION CONSIDERATIONS

8.4 INTERFACES IN ARCHIVE SYSTEMS

8.5 ARCHIVE SOLUTIONS

8.6 OPERATIONAL AND ORGANISATIONAL ASPECTS

8.7 SUMMARY AND OUTLOOK

9 Business Continuity

9.1 GENERAL CONDITIONS

9.2 STRATEGIES OF BUSINESS CONTINUITY

9.3 PARAMETERS OF BUSINESS CONTINUITY

9.4 QUALITY OF SERVICE FOR BUSINESS CONTINUITY

9.5 BUSINESS CONTINUITY SOLUTIONS

9.6 SWITCH OF OPERATIONAL LOCATION

9.7 ORGANISATIONAL ASPECTS

9.8 SUMMARY

10 Management of Storage Networks

10.1 REQUIREMENTS

10.2 CHARACTERISATION OF MANAGEMENT INTERFACES

10.3 IN-BAND MANAGEMENT

10.4 OUT-BAND MANAGEMENT

10.5 OPERATIONAL ASPECTS OF THE MANAGEMENT OF STORAGE NETWORKS

10.6 SUMMARY

11 Removable Media Management

11.1 THE SIGNIFICANCE OF REMOVABLE MEDIA

11.2 REMOVABLE MEDIA

11.3 LIBRARIES AND DRIVES

11.4 PROBLEMS AND REQUIREMENTS IN RESPECT OF REMOVABLE MEDIA MANAGEMENT

11.5 THE IEEE 1244 STANDARD FOR REMOVABLE MEDIA MANAGEMENT

11.6 SUMMARY

12 The SNIA Shared Storage Model

12.1 THE MODEL

12.2 EXAMPLES OF DISK-BASED STORAGE ARCHITECTURES

12.3 EXTENSION OF THE SNIA SHARED STORAGE MODEL TO TAPE FUNCTIONS

12.4 EXAMPLES OF TAPE-BASED BACKUP TECHNIQUES AND ARCHITECTURES

12.5 SUMMARY

13 Final Note

Glossary

Annotated Bibliography

Appendix A: Proof of Calculation of the Parity Block of RAID 4 and 5

Appendix B: Checklist for the Management of Storage Networks

B.1 APPLICATIONS

B.2 DATA

B.3 RESOURCES

B.4 NETWORK

Index

First published under the title Speichernetze, Grundlagen und Einsatz von Fibre Channel SAN, NAS, iSCSl und InfiniBand.

ISBN: 3-89864-135-X by dpunkt.verlag GmbH

Copyright © 2003 by dpunkt.verlag GmbH, Heidelberg, Germany

1st edition of the English translation first published 2004.

Translation Copyright © 2004, John Wiley & Sons Ltd.

This edition first published 2009

Translation Copyright © 2009, John Wiley & Sons Ltd.

Registered office

John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, United Kingdom

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com.

The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. 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 or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data:

Storage networks explained: basics and application of Fibre Channel SAN, NAS, iSCSI, InfiniBand, and FCoE/Ulf Troppens… [et al.]. – 2nd ed.

p. cm.

Rev. ed. of Storage networks explained/Ulf Troppens. c2004.

ISBN 978-0-470-74143-6 (cloth)

1. Storage area networks (Computer networks) 2. Information storage and retrieval systems. I. Troppens, Ulf. II. Troppens, Ulf. Storage networks explained.

TK5105.86.T78 2009

004.6–dc22

2009014224

For Silke, Hannah, Nina, and Julia

You keep showing me what really matters in life.

For Christina, Marie and Tom

For your love and support.

For Christel

Only your patience and your understanding have made my contribution to this book possible.

For Susann

In Love.

For Tineke, Daniel and Marina

For the love, motivation and reassurance you have always given me.

About the Authors

The authors are employed at IBM’s storage competence center in Mainz, Germany. They work at the interface between technology and customers. Their duties cover a wide field of responsibilities. They develop and test new software for storage networks. They present the latest hardware and software products in the field of storage networks to customers and explain their underlying concepts. Last but not least they deploy and support respective hardware and software in customer environments.

Ulf Troppens (centre) studied Computer Science at the University of Karlsruhe. Since 1989 he has been primarily involved in the development and administration of Unix systems, storage systems, data and storage networks and distributed applications.

Rainer Erkens (left) studied Mathematics at the University of Mainz. His experience in the management of computers and distributed applications goes back to 1992. Since 2005 he is a technical support manager in IBM’s European Storage Competence Center.

Wolfgang Müller-Friedt (right) studied Computer Science at the FH Darmstadt. He is a software architect focussing on the software development of management applications for storage networks which support open standards such as SMI-S and IEEE 1244.

Nils Haustein (left front) studied Electrical Engineering at the TU Chemnitz. For several years he is with IBM’s advanced technical sales support in Europe where he is focussing on digital archiving.

Rainer Wolafka (right front) studied Electrical Engineering at the FH Frankfurt and Software Engineering at the Santa Clara University. Since 1997 he is working in the field of storage networks and the software development of management applications for storage networks.

Foreword to the Second Edition by Hermann Strass

A book on the subject of storage networks is especially important during these fast-moving times. The technology for storage networking is basically bringing with it new structures and procedures that will remain topical in the foreseeable future regardless of incremental differences and changes in products. This book is based on the experience of its authors in their day-to-day work with the material. It provides system administrators and system planners in particular with the tools they need for an optimal selection and cost-effective implementation of this complex technology, the use and operation of which currently seems indispensable in view of the ever-increasing storage quantities in companies. The technology of networked storage provides demonstrable and important cost savings. Growth therefore continues even in an unfavourable economic climate.

Storage quantities are growing because we are now working much more in colour, in three-dimension and digitally than was the case years ago. Furthermore, legal regulations that exist in the European Union and in other countries make the electronic/digital storage of all business data compulsory. The law no longer allows old business documents to be filed in printed form in archives. Data quantities continue to increase in good times as well as bad. Even lost contracts and the related data must be stored digitally. The legal regulations on their own are thus ensuring that a certain amount of growth in data is inevitable.

In the past, data was stored on disk and tape drives that were connected directly to a server. Storage was operated as a peripheral to the computer. Access rights, virus protection and other functions could thus be performed on the relevant computer (server). For reasons that are explained in detail in this book, this mode of operation is no longer practical today. Storage has been detached from the servers and combined to form a separate storage network. This has resulted in a fundamentally different approach to dealing with storage. The new procedures required will continue to be developed into the near future. Data storage therefore has a value of its own. It is no longer a matter of attaching another disk drive to a server.

Today stored data and the information it contains are the crown jewels of a company. The computers (servers) needed for processing data can be purchased by the dozen or in larger quantities – individually as server blades or packed into cabinets – at any time, integrated into a LAN or a WAN or exchanged for defective units. However, if stored data is lost, restore of it is very expensive and time-consuming, assuming that all or some of it can even be recovered. As a rule, data must be available ‘around the clock’. Data networks must therefore be designed with redundancy and high availability.

These and related topics are covered in detail in this book. The approach is based upon the current state of technology only to a certain degree. What is more important is the description of the fundamental topics and how they relate to one another. This coverage goes beyond the scope of even lengthy magazine articles and will continue to be topical in the future. This is the only book available in the market today that covers this subject so comprehensively.

The requirements of storage networks are fundamentally different from those of the familiar local networks (LANs). Storage networks have therefore almost exclusively been using Fibre Channel technology, which was specially developed as a connection technology for company-critical applications. Storage networking is not a short-term trend and efforts are therefore currently underway to use other existing (for example, Ethernet-LAN-TCP/IP) network technologies as well as new ones that are coming on the market (for example InfiniBand and FCoE). Under certain circumstances these are totally sensible alternatives. This book highlights which selection criteria play a role here. It is usually not technical details or prejudices that are decisive but rather usage requirements, existing infrastructure and devices, along with a careful assessment of the future development in companies. The aim of this book is to provide valuable help in structural planning and the selection of devices and software.

The importance of networked storage technology has grown substantially since the first edition was printed. For the reasons mentioned in this book and due to regulatory requirements, even medium-sized companies need to manage large quantities of data and make them available for many years. This is why the sections on storage archiving have been considerably expanded in the new edition of this book. In a global economy business continuity is overly important for survival. This second edition devotes extensive coverage to this topic.

Overall this book is an excellent work. It explains the chosen subject comprehensively and in great detail, based on solid technical foundations. It is hoped that it will gain a wide circulation, particularly as it corrects a great many half-truths with its presentation of facts and addresses the usual prejudices.

Hermann Strass

Preface by the Authors

This Preface answers the following main questions:

What does this book deal with?Who should read this book?How should this book be read?Who has written this book?

WHAT DOES THIS BOOK DEAL WITH?

The technology of storage networks fundamentally changes the architecture of IT systems. In conventional IT systems, storage devices are connected to servers by means of SCSI cables. The idea behind storage networks is that these SCSI cables are replaced by a network, which is installed in addition to the existing LAN. Server and storage devices can exchange data over this new network using the SCSI protocol. Storage networks have long been a known quantity in the world of mainframes. Fibre Channel, iSCSI, FCoE and Network Attached Storage (NAS) are now also taking storage networks into the field of Open Systems (Unix, Windows, OS/400, Novell Netware, MacOS).

Storage networks are a basic technology like databases and LANs. Storage was previously installed in the servers. Now most storage capacity is provided in external devices that are linked to servers over a storage network. As a result, anyone who is involved in the planning or operation of IT systems requires basic knowledge about the fundamentals and the use of storage networks. These networks are almost as widespread as SCSI, SAS and SATA but are more complex than LANs and TCP/IP.

The book is divided into two parts. Part I deals with fundamental technologies relating to storage networks. It guides the reader from the structure and operating method of storage devices through I/O techniques and I/O protocols to the file systems and storage virtualisation.

The second part of this book presents applications that utilise the new functions of storage networks and intelligent disk subsystems. The emphasis here is on the shared use of resources that are available over a storage network, scalable and adaptable storage architectures, network backup and digital archiving. Another important focus of the book is business continuity with strategies for continuous and loss-free operation as protection against small failures and large catastrophes. Further focal points are the discussions on the management of storage networks and the management of removable media. Last but not least, the SNIA Shared Storage Model provides a reference model to describe storage networks.

At the end of the book we have added a glossary, an index and an annotated bibliography, which in addition to further literature also highlights numerous freely available sources on the Internet.

Section 1.4 sets out in detail the structure of the book and the relationships between the individual chapters. Figure 1.7 illustrates the structure of the book. At this point, it is worth casting a glance at this illustration. Note that the illustration also describes the subjects that we will not be covering.

Long before the second edition was printed, many readers of the first edition wanted to know what the differences are between the two editions. Here we want to express that our approach was successful, we aimed at introducing basic concepts rather than presenting actual products and overly technical details. The chapter on I/O techniques was the only one that required some updating on Fibre Channel and iSCSI. The key distinction of the second edition is the addition of two new chapters covering the topics of digital archiving and business continuity. We have also expanded the coverage on the copy services of intelligent disk subsystems.

WHO SHOULD READ THIS BOOK?

Our approach is, first, to explain the basic techniques behind storage networks and, secondly, to show how these new techniques help to overcome problems in current IT systems. The book is equally suitable for beginners with basic IT knowledge and for old hands. It is more an introduction to the basic concepts and techniques than a technical reference work. The target group thus includes:

System administrators and system architectsSystem consultantsDecision makersUsersStudents

After reading the whole book you will be familiar with the following:

The concepts of storage networks and their basic techniquesUsage options for storage networksProposed solutions for the support of business processes with the aid of storage networksThe advantages of storage networksNew possibilities opened up by storage networks.

HOW SHOULD THIS BOOK BE READ?

There are two options for reading this book. Those readers who are only interested in the concepts and usage options of storage networks should read Chapter 1 (Introduction) and Part II (Application and Management of Storage Networks); they can use Part I as a reference to look up any basic technical information they might require. Readers who are also interested in the technical background of storage networks should read the book through from the beginning.

WHO HAS WRITTEN THIS BOOK?

Ulf Troppens began work on this book in 2001. Rainer Erkens joined him soon after, providing his contributions on the topics of storage virtualisation, management of storage networks and NDMP for the first edition in 2002. In 2004 Wolfgang Müller-Friedt expanded the English translation – which was presented with the ‘Editor’s Choice Award 2005’ by Linux Journal – with his sound knowledge of magnetic tape, tape libraries and their management. Lastly, the second edition has been expanded considerably through contributions by Nils Haustein (digital archiving) and Rainer Wolafka (business continuity).

All five authors have different roles at the Storage Competence Center of IBM in Mainz, Germany. Our responsibilities range from the development and testing of new software for storage networks to providing guidance to customers on the procurement of suitable products and the respective underlying concepts as well as on the installation and support of relevant hardware and software for customer environments. We advise customers on how storage networks can help to solve problems in their current IT systems. This experience has made us familiar with the types of questions customers have in respect of storage networks. Our involvement extends to customers with experience in storage networks as well as to those who are novices in this field. The positive feedback we have received from readers of the first edition show that our work has helped us to structure the content of this book and to choose topics in a way that are important to readers of books on storage networks.

Our intention has been to take off our ‘IBM hats’ and to write this book from an unbiased viewpoint. As employees of IBM in the area of storage technology, the experience and opinions that have been formed in our day-to-day work have of course had some influence on this book. In this connection, we have to be very familiar with our own company’s products as well as with those of our competitors and to position these products so that we inevitably have a view that goes beyond the IBM scope. In the end, this book is our personal work and has no connection with IBM apart from our employee relationship. Most importantly, this book does not represent any of the official opinions of IBM.

ACKNOWLEDGEMENTS FOR THE SECOND EDITION

We would like to give special thanks to our technical advisors on the second edition: Dirk Jahn (Archiving), Hans-Peter Kemptner (Business Continuity), Robert Haas (Limitations of RAID 5) and Hermann Strass for the Foreword. Other contributions were made by Jens-Peter Akelbein. We also appreciate the help we received on the publishing side from Rene Wiegand (copy-editing), Ulrich Kilian (LaTeX) and Rene Schoenfeld (editorial), all who helped to make our manuscript ready for printing.

With regard to the second English edition we would like to thank Birgit Gruber, Tiina Ruonamaa, Brett Wells, Liz Benson, Anna Smart, Sarah Tilley, Mary Lawrence and Sarah Hinton (all Wiley & Sons) as well as Deepthi Unni and her team at Laserwords. Last but not the least we thank Hedy Jourdan for the great translation of the new parts from German to English.

ACKNOWLEDGEMENTS FOR THE FIRST EDITION

We would also like to use this preface to thank some of the people who have made a significant contribution to the first edition of this book. From a chronological point of view, we should start by mentioning the editorial department of iX magazine and the copy-editing staff of dpunkt.verlag as they set the whole project in motion in March 2001 with the question ‘Could you see yourselves writing a book on the subject of storage in the network?’

Regarding content, our colleagues from the IBM Mainz storage community, especially the former SAN Lab and the current TotalStorage Interoperability Center (meanwhile renamed to Systems Lab Europe), deserve mention: Without the collaboration on storage hardware and software with customers and employees of partner companies, business partners and IBM, and without the associated knowledge exchange, we would lack the experience and knowledge that we have been able to put into this book. The list of people in question is much too long for us to include it here. The cooperation of one of the authors with the students of the BAITI 2000 course of the Berufsakademie Mannheim (University of Applied Science Mannheim), from whom we have learnt that we have to explain subjects such as ‘RAID’, ‘disk subsystems’, ‘instant copy’, ‘remote mirroring’ and ‘file server’, was also valuable from a didactic point of view.

With regard to quality control, we thank our proofreaders Axel Köster, Bernd Blaudow, Birgit Bäuerlein, Frank Krämer, Gaetano Bisaz, Hermann Strass, Jürgen Deicke, Julia Neumann, Michael Lindner, Michael Riepe, Peter Münch, René Schönfeldt, Steffen Fischer, Susanne Nolte, Thorsten Schäfer, Uwe Harms and Willi Gardt, as well as our helpers at dpunkt.verlag, whose names we do not know.

We should emphasise in particular the many constructive suggestions for improvement by Susanne Nolte, who also contributed a few paragraphs on ‘DAFS’, and the numerous comments from our colleagues Axel Köster and Jürgen Deicke and our manuscript reader René Schönfeldt. In this connection, the efforts of Jürgen Deicke and Tom Clark should also be mentioned regarding the ‘SNIA Recommended Reading’ logo, which is printed on the front cover of the book.

With regard to the first English edition of this book we have to thank even more people: First of all, we would like to thank René Schönfeldt from dpunkt.verlag for convincing Birgit Gruber from Wiley & Sons to invest in the translation. We greatly appreciate Birgit Gruber for taking a risk on the translation project and having so much patience with all our editorial changes. Rachel Waddington did an outstanding job of translating the text and all the figures from German into English. Last but not least, we would like to thank Daniel Gill for leading the production process, including copy-editing and typesetting, and we would like to thank the team at Laserwords for typesetting the whole book.

Closing comments

Finally, the support of our parents, parents-in-law and partners deserves mention. I, Nils Haustein, would like to thank my dear wife Susann who gave me a lot of ‘computer time’ and the opportunity to make a contribution to this book. I, Rainer Wolafka, would like to thank my dear wife Tineke for her support and her constant encouragement and motivation to work on this book and to my son Daniel for understanding why I did not always have the time he deserved during this time. I, Wolfgang Müller-Friedt, would like to thank my dear wife Christel for her patience, her emotional support and for many more reasons than there is room to list in these notes. I, Ulf Troppens, at this point would like to thank my dear wife Silke for her support and for taking many household and family duties off my hands and thus giving me the time I needed to write this book. And I, Rainer Erkens, would like to thank my dear partner Christina, who never lost sight of worldly things and thus enabled me to travel untroubled through the world of storage networks, for her support. We are pleased that we again have more time for children, our families and friends. May we have many more happy and healthy years together.

Mainz, April 2009

Ulf Troppens

Rainer Erkens

Wolfgang Müller-Friedt

Nils Haustein

Rainer Wolafka

List of Figures and Tables

FIGURES

1.1 Server-Centric IT architecture 1.2 Inflexible allocation of free storage capacity 1.3 Storage-Centric IT architecture 1.4 Server upgrade: preparation of a new server 1.5 Server upgrade: testing a new server 1.6 Server upgrade: putting a new server into service 1.7 Structure of this book 2.1 Connection of servers to a disk subsystem 2.2 Architecture of intelligent disk subsystems 2.3 Flexible allocation of free storage capacity 2.4 Comparison: large and small internal hard disks 2.5 Internal I/O channels: active and active/passive 2.6 Internal I/O channels: active/active 2.7 Virtualisation by RAID 2.8 Replacing a defective physical hard disk 2.9 RAID 0 (striping) 2.10 RAID 1 (mirroring) 2.11 RAID 0+1 (mirrored stripes) 2.12 RAID 10 (striped mirrors) 2.13 RAID 0+1: failure of a physical hard disk 2.14 RAID 10: failure of a physical hard disk 2.15 RAID 4 (parity disk) 2.16 RAID 4 and RAID 5: write penalty 2.17 RAID 5 (striped parity) 2.18 Instant copy 2.19 Space-efficient instant copy 2.20 High availability with remote mirroring 2.21 Data flow of synchronous remote mirroring 2.22 Data flow of asynchronous remote mirroring 2.23 Synchronous and asynchronous remote mirroring combined 2.24 Consistency groups in heterogeneous environments 2.25 Write order consistency for asynchronous remote mirroring 2.26 Chaos without LUN masking 2.27 Order with LUN masking 3.1 The physical I/O path from the CPU to the storage system 3.2 The physical I/O path within a disk subsystem 3.3 Connection of storage over SCSI daisy chain 3.4 SCSI target IDs and SCSI LUNs 3.5 Priority of the SCSI target IDs 3.6 Twin-tailed SCSI cabling 3.7 Small SCSI SAN with multiport storage systems 3.8 The Fibre Channel protocol stack 3.9 The Fibre Channel topologies: point-to-point, arbitrated loop, fabric 3.10 Different plug types for fiber-optic cables 3.11 NRZ and Manchester encoding 3.12 Jitter 3.13 Fibre Channel exchange, sequence and frame 3.14 The Fibre Channel frame format 3.15 Link flow control and end-to-end flow control 3.16 Fibre Channel Class 2: error-free transmission 3.17 Fibre Channel Class 2: transmission error 3.18 Fibre Channel Class 3: error-free transmission 3.19 Fibre Channel Class 3: transmission error 3.20 Fabric login, N-Port login and process login 3.21 Addressing in the fabric 3.22 Fibre Channel application protocols based on the example of FCP 3.23 Latency of Fibre Channel components 3.24 Multiple bandwidths in the fabric topology 3.25 Bottleneck: inter switch link (ISL) 3.26 Fibre Channel arbitrated loop 3.27 Cascading of Fibre Channel arbitrated loops 3.28 Public loop and private loop 3.29 Communication between arbitrated loop and fabric 3.30 Fibre Channel-to-SCSI bridges 3.31 Typical entry-level configuration: dual fabric 3.32 Fibre Channel link extender 3.33 Today: Fibre Channel islands 3.34 Tomorrow: one large InterSAN? 3.35 Connecting storage via iSCSI 3.36 Booting via iSCSI 3.37 Gateway protocol: internet FCP (iFCP) 3.38 Tunnelling protocol: Fibre Channel over IP (FCIP) 3.39 Comparison of the protocol stacks of FCP, FCIP, iFCP and iSCSI 3.40 Comparison of the frame formats of FCP, FCIP, iFCP and iSCSI 3.41 A joint network for LAN, MAN, WAN and SAN 3.42 CPU loading: TCP/IP/Ethernet versus Fibre Channel 3.43 Migration path from Fibre Channel to iSCSI 3.44 Architecture of InfiniBand 3.45 Example scenario of an InfiniBand interconnection 3.46 Virtual Interface Architecture (VIA) 3.47 I/O consolidation 3.48 FCoE protocol stack 3.49 FCoE frame format 3.50 FCoE end device (FCoE Node, ENode) 3.51 FCoE switch 3.52 Virtual link 3.53 FCoE Initialization Protocol (FIP) 3.54 FCoE SAN 3.55 Integration of FCoE SAN and Fibre Channel 3.56 FCoE SAN with blade server 3.57 Enhanced Transmission Selection (ETS) 4.1 File system and volume manager 4.2 File systems and databases 4.3 Volume manager 4.4 RAID in the volume manager 4.5 Network file system 4.6 NAS server and NAS gateway 4.7 Eye of the needle: file server 4.8 DAFS cluster 4.9 DAFS data flow 4.10 A comparison between NFS, uDAFS and fDAFS 4.11 Shared disk file systems 4.12 GPFS architecture 4.13 GPFS cluster, GPFS node set and GPFS file system 4.14 GPFS token management for cache synchronisation 4.15 Comparison of the I/O paths: SCSI, Fibre Channel, FCoE, iSCSI, NAS 4.16 Storage networks with NAS servers 5.1 Basic idea: virtualisation in the storage network 5.2 Virtualisation locations in the I/O path 5.3 Virtualisation in the disk subsystem 5.4 Virtualisation in the storage network 5.5 Virtualisation in the host bus adapter 5.6 Virtualisation in the volume manager 5.7 Mirroring in the volume manager for high fault-tolerance 5.8 Striping in the volume manager for high write performance 5.9 Limitations of storage sharing 5.10 Limitations due to incompatible device drivers 5.11 Storage virtualisation as an abstraction layer 5.12 Virtualisation on block level 5.13 Virtualisation on file level 5.14 Virtualisation on different levels of the storage network 5.15 Symmetric storage virtualisation 5.16 Structure of the metadata controller in symmetric virtualisation 5.17 Asymmetric storage virtualisation 5.18 Structure of the metadata controller in asymmetric virtualisation 6.1 Layering of the transmission techniques 6.2 Storage networks in the I/O path 6.3 Storage networks within disk subsystems 6.4 Inflexible: storage assignment in server-centric systems 6.5 Better: storage pooling in storage-centric IT systems 6.6 Partitioning of a tape library 6.7 Tape library sharing 6.8 Real-time data sharing 6.9 Partitioning of the data set 6.10 Parallel databases and file systems 6.11 Protection against the failure of an I/O bus: redundant SCSI cable 6.12 Protection against the failure of an I/O bus: dual storage network 6.13 Problem: operating system recognises hard disk several times 6.14 Solution: multipathing software 6.15 Protection against the failure of a server: server cluster 6.16 Case study: cluster for file servers – failure of a server 6.17 Server cluster and redundant I/O buses 6.18 Protection against the failure of a disk subsystem: remote mirroring 6.19 Protection against the failure of a disk subsystem: volume manager mirroring 6.20 Volume manager mirroring with server cluster and redundant I/O buses 6.21 The logical I/O path for volume manager mirroring with redundant I/O buses 6.22 Case study: protection of an important database 6.23 Case study: protection of an important database (failure of the primary data centre) 6.24 Shared-null configuration 6.25 Shared-null configuration (failure of a server) 6.26 Shared-nothing cluster 6.27 Shared-nothing cluster (failure of a server) 6.28 Enhanced shared-nothing cluster 6.29 Enhanced shared-nothing cluster (load balancing) 6.30 Shared-everything cluster 6.31 Shared-everything cluster (failure of a server) 6.32 Three tier architecture 6.33 Three tier architecture (implementation) 6.34 Web architecture (five tier architecture) 6.35 Web architecture (implementation) 6.36 Web architecture based upon the ‘travel portal’ case study 6.37 Extension of the travel portal 6.38 Extension of the travel portal: cache server 7.1 Example environment for network backup 7.2 Storage hierarchy in the backup server 7.3 Performance bottlenecks for traditional network backup 7.4 Separate LAN for network backup 7.5 Multiple backup servers 7.6 Backup server and application server on the same computer 7.7 Server-free backup 7.8 LAN-free backup 7.9 LAN-free backup with shared disk file system 7.10 Application server-free backup 7.11 Tape library sharing 7.12 Backup of a NAS server by means of a network backup system 7.13 Architecture of the Network Data Management Protocol (NDMP) 7.14 NDMP: Backup to a local tape drive 7.15 NDMP: Backup to a remote tape drive 7.16 NDMP: LAN-free backup with NDMP 7.17 Translator services in NDMP Version 5 7.18 Architecture of databases 7.19 Backup of archive log files 8.1 Reference architecture for digital archive systems 8.2 General infrastructure for network backup 8.3 General infrastructure for digital archiving 8.4 General infrastructure for backing up archive data 8.5 Life cycle of classical data and information 8.6 Life cycle of archive data 8.7 Example: multilevel guarantee of data integrity 8.8 Data mirroring in the archive storage 8.9 Hierarchical storage management (HSM) for archive data 8.10 Interfaces in the reference architecture for archive systems 8.11 Protocol converters for incompatible interfaces 8.12 Interface between multiple DMS 8.13 Standardised interfaces for archive systems 8.14 Archiving of emails using archive storage 8.15 Archiving of emails using archive storage and DMS 8.16 Archiving of files using archive storage and file server emulation 8.17 Archiving of files using archive storage and protocol converter 8.18 Archiving of files using archive storage and DMS 8.19 Archiving of ERP data using archive storage 8.20 Archiving of ERP data using archive storage and DMS 8.21 Archiving of PACS or HIS data using archive storage 8.22 Central archiving of PACS and HIS data using DMS 8.23 Central archiving using one DMS 8.24 Central archiving using federated DMS 9.1 Phases and dependencies of system restart 9.2 Business continuity for the web architecture 9.3 The cost–time window of business continuity 9.4 Cycle of risk management 9.5 Strategies for business continuity 9.6 Parameters of availability: MTTR, MTTF and MTBF 9.7 Availability of sequentially coupled components 9.8 Availability of parallel-coupled components 9.9 Parameter of business continuity: RTO 9.10 Parameter of business continuity: RPO 9.11 The seven-tier model for business continuity 9.12 Solution segments of the seven-tier model 9.13 D2D2T for rapid restore 9.14 RPO for synchronous and asynchronous remote mirroring 9.15 Remote mirroring combined with instant copy for disaster recovery 9.16 Volume manager mirroring combined with remote mirroring for high availability and disaster recovery 9.17 Simplification without volume manager mirroring 9.18 Remote mirroring with instant copy for rolling disasters 9.19 Protection against rolling disasters in backup data centres 9.20 Combination of high availability and disaster recovery 10.1 Devices in the storage network 10.2 In-band management 10.3 Out-band management 10.4 SNMP architecture 10.5 MIB hierarchy 10.6 Web Based Enterprise Management (WBEM) 10.7 The three columns of WBEM 10.8 CIM: relationships and inheritance of attributes 10.9 CIM schemas 10.10 WBEM architecture 10.11 SMI-S architecture 10.12 DMI architecture 11.1 Independent versus attached media changer 11.2 A tape library with proprietary interfaces 11.3 Central removable media management 11.4 Drive sharing with mount request queues 11.5 Scratch pools spanning multiple libraries 11.6 Drive pools spanning multiple libraries 11.7 Integration of offline storage (vaulting) 11.8 Cartridge life cycle 11.9 Architecture of the IEEE standard 1244 12.1 Main components of the SNIA shared storage model 12.2 The layers of SNIA shared storage model 12.3 Implementation of the file/record layer 12.4 Functionality of the block aggregation layer 12.5 Combination of block and file layer 12.6 Access paths 12.7 Caching 12.8 Clustering 12.9 Resource and data sharing 12.10 Service subsystem 12.11 Example – direct attached block storage 12.12 Example – storage network attached block storage 12.13 Example – block storage aggregation in a storage device: SAN appliance 12.14 Example – network attached block storage with metadata server: asymmetric block service 12.15 Example – multi-site block storage 12.16 Example – file server 12.17 Example – NAS head 12.18 Example – file server metadata manager 12.19 Example – object-based storage device (OSD) 12.20 Tape modelling (logical) 12.21 Tape modelling (physical) 12.22 Tape extension of the SNIA shared storage model 12.23 Example – file backup (tar) 12.24 Example – file system volume backup (dump) 12.25 Example – volume backup (dd) 12.26 Example – file backup to virtual tape 12.27 Example – direct attached tape 12.28 Example – LAN attached tape 12.29 Example – shared tape drive 12.30 Example – partitioned tape library 12.31 Example – virtual tape controller 12.32 Example – virtual tape controller with disk cache 12.33 Example – data mover for tape 12.34 Example - file server with tape drive 12.35 Example – file server with external tape 12.36 Example – file server with data mover

TABLES

2.1 Failure probability of a RAID 5 array 2.2 Comparison of RAID levels 3.1 SCSI cable lengths and transmission rates 3.2 Fibre Channel roadmap 3.3 Well-defined Fibre Channel addresses 3.4 Comparison of iSCSI, iFCP, FCIP and FCP 4.1 Comparison of snapshot and instant copy 4.2 Comparison of Fibre Channel, FCoE, iSCSI and NAS 8.1 Comparison of backup and archiving 8.2 Comparison of the three basic WORM technologies 9.1 Risk factors for IT operations 9.2 Availability and outage times for a 24×7 operation 9.3 RTO and RPO for high availability and disaster recovery 9.4 The seven-tier model for business continuity 11.1 The IEEE 1244 data model 12.1 Data – container relationships A.1 Calculation of parity block for RAID 4 and RAID 5

1

Introduction

The purpose of this chapter is to convey the basic idea underlying this book. To this end we will first describe conventional server-centric IT architecture and sketch out its limitations (Section 1.1). We will then introduce the alternative approach of storage-centric IT architecture (Section 1.2), explaining its advantages using the case study ‘Replacing a Server with Storage Networks’ (Section 1.3). Finally, we explain the structure of the entire book and discuss which subjects are not covered (Section 1.4).

1.1 SERVER-CENTRIC IT ARCHITECTURE AND ITS LIMITATIONS

In conventional IT architectures, storage devices are normally only connected to a single server (Figure 1.1). To increase fault tolerance, storage devices are sometimes connected to two servers, with only one server actually able to use the storage device at any one time. In both cases, the storage device exists only in relation to the server to which it is connected. Other servers cannot directly access the data; they always have to go through the server that is connected to the storage device. This conventional IT architecture is therefore called server-centric IT architecture. In this approach, servers and storage devices are generally connected together by SCSI cables.