`
`|PR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`DHPN-1003 I Page 1 of 181
`DHPN-1003 I Page 1 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Clusters for
`High Availability
`
`DHPN-1003 I Page 2 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`•
`
`Hewlett-Packard Professional Books
`
`Blinn
`
`Blommers
`Costa
`
`Crane
`
`Fernandez
`Fristrup
`
`Fristrup
`Grady
`
`Grosvenor, Ichiro,
`O'Brien
`
`Gunn
`Helsel
`Helsel
`Kane
`Knouse
`Lewis
`Madell, Parsons, Abegg
`Malan, Letsinger,
`Coleman
`McFarland
`
`Pmtable Shell Programming: An Extensive Collection of
`Bourne Shell Examples
`
`Practical Planning for Network Growth
`
`Planning and Designing High Speed Networks
`Using lOOVG-AnyLAN, Second Edition
`
`A Simplified Approach to Image Processing: Classical and
`Modem Techniques
`
`Configuring the Common Desktop Environment
`
`USENET: Netnews for Everyone
`
`The Essential Web Smfer Survival Guide
`
`Practical Software Metrics for Project
`Management and Process Improvement
`
`Mainframe Downsizing to Upsize Your Business:
`IT-Preneuring
`
`A Guide to Net Ware® for UNIX®
`Graphical Programming: A Tutorial for HP VEE
`
`Visual Programming with HP-VEE
`
`PA-RISC 2.0 Architecture
`
`Practical DCE Programming
`
`The Art & Science of Smalltalk
`
`Developing and Localizing International Software
`
`Object-Oriented Development at Work: Fusion In
`the Real World
`
`X Wrndows on the World: Developing
`Internationalized Software with X, Motu®, and CDE
`
`McMinds/Whitty
`
`Writing Your Own OSF/MotifWidgets
`
`Phaal
`Poniatowski
`Poniatowski
`
`Thomas
`
`Weygant
`
`Witte
`
`LAN Traffic Management
`
`The HP-UX System Administrator's "How To" Book
`
`HP-UX lO.x System Administration "How To" Book
`
`Cable Television Proof-of-Performance: A Practical
`Guide to Cable TV Compliance Measurements Using
`a Spectrum Analyzer.
`
`Clusters for High Availability: A Primer ofHP-UX Solutions
`
`Electronic Test Instruments
`
`DHPN-1003 I Page 3 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Clus~ers for High
`Availability
`
`A Primer of HP-UX Solutions
`
`Peter Weygant
`
`Hewlett-Packard Company
`
`Prentice Hall PTR
`Upper Saddle River, New Jersey 07458
`
`DHPN-1003 I Page 4 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Editorial/Production Supervision: Jomme Anzalone
`Acquisitions Editor: Karen Gettman
`Manufacturing Manager: Alexis R. Heydt
`Cover Design: Design Source
`Manager, Hewlett-Packard Press: Pat Pekary
`
`© 1996 by Hewlett-Packard Company
`
`It Published by Prentice Hall PTR
`
`Prentice-Hall, Inc.
`A Simon & Schuster Company
`Upper Saddle River, NJ 07458
`
`All rights reserved. No part of this book may be
`reproduced, in any form or by any means, without
`permission in writing from the publisher.
`
`MC/ServiceGuard and MC/LockManager are registered trademarks of Hewlett-Packard
`Company. Oracle is a trademark of Oracle Corporation. Symmetrix and EMC are
`trademarks of EMC Corporation. NFS is a trademark of Sun Microsystems, Inc. UNIX
`is a registered trademark in the United States and in other countries, licensed exclusively
`through X/ Open Company, Ltd.
`The publisher offers discounts on this book when ordered in bulk quantities.
`For more information, contact the Corporate Sales Department, PTR Prentice Hall, One
`Lake Street, Upper Saddle River, NJ 07458. Phone: 800-382-3419. FAX: 201-236-7141. e-mail:
`corpsales@prenhall.com
`
`Printed in the United States of America
`
`10 9 8 7
`
`ISBN 0-13-494758-4
`
`HP Part Number 83936-90007
`
`Prentice-Hall International (UK) Limited, London
`Prentice-Hall of Australia Ply. Limited, Sydney
`Prentice-Hall of Canada, Inc., Toronto
`Prentice-Hall Hispanomnericana S.A., Mexico
`Prentice-Hall of India Private Limited, New Delhi
`Prentice-Hall of Japan, Inc., Tokyo
`Simon & Schuster Asia Pte. Ltd., Singapore
`Editora Prentice-Hall do Brasil, Ltd., Rio de Janeiro
`
`DHPN-1003 I Page 5 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Contents
`
`Foreword
`
`Preface
`
`Acknowledgements
`
`About the Author
`
`1
`
`BASIC HIGH AVAILABILITY CONCEPTS
`
`What is High Availability?
`Available
`Highly Available
`Highly Available Computing
`Service Levels
`Continuous Availability
`Fault Tolerance
`Matching Availability to User Needs
`Choosing a Solution
`
`High Availability as a Business Requirement
`High Availability as Insurance
`
`v
`
`xiii
`
`XV
`
`xvi
`
`xvii
`
`1
`
`2
`2
`3
`4
`5
`6
`6
`7
`7
`
`8
`
`8
`
`DHPN-1003 I Page 6 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Contents
`
`High Availability as Opportunity
`9
`Cost of High Availability
`10
`What Are the Measures of High Availability?
`11
`Calculating Availability
`11
`Expected Period of Operation
`12
`Calculating Mean Time Between Failures
`14
`Understanding the Obstacles to High Availability 16
`Duration of Outages
`17
`Time Lines for Outages
`18
`Causes of Planned Downtime
`20
`Causes of Unplanned Downtime
`22
`Severity of Unplanned Outages
`23
`Designing for Reaction to Failure
`23
`Identifying Points of Failure
`24
`Preparing Your Organization for High
`Availability
`25
`Stating Availability Goals
`25
`Building the Appropriate Physical Environment
`27
`Creating Automated Processes
`27
`Using a Development and Test Environment
`28
`Maintaining a Stock of Spare Parts
`28
`Defining an Escalation Process
`29
`Planning for Disasters
`29
`Training System Administration Staff
`29
`Using Dry Runs
`30
`Documenting Every Detail
`30
`The Starting Point for a Highly Available System 31
`Basic Hardware Reliability
`31
`Software Quality
`32
`Intelligent Diagnostics
`32
`
`vi
`
`DHPN-1003 I Page 7 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Contents
`
`Comprehensive System Management Tools
`Maintenance and Support Services
`
`Moving to High Availability
`
`Summary
`
`2
`
`CREATING A HIGH AVAILABILITY CLUSTER
`
`Identifying Single Points of Failure in a
`Stand-alone System
`
`33
`33
`34
`35
`
`39
`
`40
`
`Eliminating Power Sources as Single
`Points of Failure
`Individual UPS Units
`Power Passthrough UPS Units
`
`45
`45
`46
`48
`Eliminating Disks as Single Points of Failure
`Data Protection with Disk Arrays
`49
`Data Protection with Software Mirroring
`51
`Eliminating the SPU as a Single Point of Failure 54
`Eliminating Single Points of Failure in Networks 57
`Points of Failure in Client Connectivity
`57
`Examples of Points of Failure
`58
`Points of Failure in Inter-Node Communication
`60
`Eliminating the Failure Points
`60
`Providing Redundant LA.N Connections
`61
`Configuring Local Switching of LAN Interfaces
`61
`Providing Redundant FDDI Connections
`66
`Using Dual Attached FDDI
`68
`Redundancy for Dialup Lines, Hardwired Serial Connec-
`tions and X.25
`
`69
`
`vii
`
`DHPN-1003/ Page 8 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Contents
`
`Eliminating Software as a Single Point of Failure 70
`Tailoring Applications for Cluster Use
`71
`
`Implementing the
`High Availability Cluster
`Complete High Availability Solution
`
`3
`
`HP'S HIGH AVAILABILITY CLUSTER
`COMPONENTS
`
`Choosing HA Architectures and Cluster
`Components
`Active/Standby Configurations
`Using MC/ServiceGuard
`Active/ Active Configurations
`Using MC/ServiceGuard
`How MCfServiceGuard Works
`Parallel Database Configuration
`Using MC/LockManager
`Oracle Parallel Server
`How MC/LockManager Works with OPS
`
`Selecting Other HA Subsystems
`Mirror Disk/UX
`High Availability Disk Storage Enclosure
`High Availability Disk Arrays
`EMC Disk Arrays
`Journaled File System
`OnLineJFS
`Transaction Processing Monitors
`Uninterruptible Power Supplies
`System and Network Management Tools
`
`viii
`
`73
`74
`
`77
`
`78
`
`79
`
`82
`84
`
`90
`91
`92
`
`94
`95
`95
`96
`97
`97
`98
`99
`99
`100
`
`DHPN-1003 I Page 9 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Contents
`
`Using Mission Critical Consulting and
`Support Services
`Availability Management Service
`Business Continuity Support
`Business Recoven; Services
`
`4
`
`SAMPLE HIGH A VAl LABILITY SOLUTIONS
`
`Highly Available NFS System for Publishing
`High Availability Software and Packages
`Hardware Configuration
`Responses to Failures
`Stock Quotation Service
`High Availability Software and Packages
`Hardware Configuration
`Responses to Failures
`Order Entry and Catalog Application
`High Availability Software and Packages
`Hardware Configuration
`Responses to Failures
`Insurance Company Database
`Two-Node OPS Configuration
`
`106
`106
`107
`109
`
`111
`
`112
`112
`114
`115
`120
`121
`122
`124
`127
`127
`131
`133
`134
`135
`
`ix
`
`DHPN-1003 I Page 10 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Contents
`
`5
`
`GLOSSARY OF HIGH AVAILABILITY
`TERMINOLOGY
`
`139
`
`AdminCenter
`Adoptive Node
`ADT
`AFR
`Alternate Node
`Annualized Failure Rate
`Architecture for HA
`Availability
`Average Downtime
`Cluster
`Cluster View
`Continuous Availability
`Custody
`Downtime
`Failure
`Failover
`Fault Tolerance
`Grouped Net
`Hardware Mirroring
`Highly Available
`Hot Plug Capability
`Hot Swap Capability
`LAN
`LAN interface
`Logical Volume Manager
`MC/LockManager
`MC/ServiceGuard
`Mean Time Between Failures
`Mean Time to Repair
`MirrorDisk/UX
`Mirroring
`
`X
`
`140
`140
`140
`140
`140
`141
`142
`142
`143
`144
`145
`145
`145
`145
`145
`146
`146
`146
`146
`146
`147
`147
`147
`147
`148
`148
`148
`148
`149
`150
`151
`
`'
`DHPN-1003/ Page 11 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Contents
`
`MTBF
`MTTR
`Network Node Manager
`Node
`Open View
`OperationsCenter
`Planned Downtime
`Primary Node
`Package
`Process Resource Manager
`RAID
`Redundancy
`Reliability
`Relocatable IP Address
`Service
`Service Level Agreement
`Shared Logical Volume Manager
`Single Point of Failure
`SLVM
`Software Mirroring
`SPOF
`SPU
`Subnet
`SwitchOverfUX
`System Processor Unit
`Transfer of Packages
`Unplanned Downtime
`Volume Group
`Index
`
`xi
`
`151
`151
`151
`151
`152
`152
`152
`152
`153
`153
`153
`154
`154
`154
`154
`155
`155
`155
`156
`156
`156
`156
`156
`156
`157
`157
`157
`157
`159
`
`DHPN-1003 I Page 12 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Foreword
`
`Foreword
`
`Over the last ten years, UNIX systems have moved from the spe(cid:173)
`cialized role of providing desktop computing power for engineers
`into the broader arena of commercial computing. This evolution is
`the result of continual dramatic improvements in functionality, re(cid:173)
`liability, performance, and supportability. We are now well into
`the next phase of the UNIX evolution: providing solutions for mis(cid:173)
`sion critical computing.
`
`To best meet the requirements of the data center for availability,
`scalability, and flexibility, Hewlett-Packard has developed a ro(cid:173)
`bust cluster architecture for HP-UX that combines multiple sys(cid:173)
`tems into a high availability cluster. Individual computers, known
`as nodes, are connected in a loosely-coupled manner, each main(cid:173)
`taining its own separate processors, memory, operating system,
`and storage devices. Special system processes bind these nodes to(cid:173)
`gether and allow them to cooperate to provide outstanding levels
`of availability and flexibility for supporting mission critical appli(cid:173)
`cations. The nodes in a cluster can be configured either to share
`data on a set of disks or to obtain exclusive access to data.
`
`To maintain Hewlett-Packard's commitment to the principles of
`open systems, our high availability clusters use standards-based
`hardware components such as SCSI disks and Ethernet LANs.
`There are no proprietary APis that force vendor lock-in, and most
`applications will run on a high availability cluster without modifi(cid:173)
`cation.
`
`xiii
`
`DHPN-1003 I Page 13 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`As the world's leading vendor of open systems, Hewlett-Packard
`is especially proud to publish this primer on cluster solutions for
`high availability. Peter Weygant has done a fine job of presenting
`the basic concepts, architectures, and terminology used in HP's
`cluster solutions. This is the place to begin your exploration of the
`world of high availability clusters.
`
`Xuan Bui
`Hewlett-Packard General Systems Division
`Research and Development Laboratory Manager
`
`xiv
`
`DHPN-1003/ Page 14 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Preface
`
`liM
`
`Et :q:iiki~ - - - - - - - - - - -- - - - -- -
`
`Preface
`
`This guide is about high availability (HA) computing through enterprise
`clusters. It presents basic concepts and terms, then describes the use of
`cluster technology to provide highly available open systems solutions
`for the commercial enterprise. Here are the topics:
`
`• Chapter 1, "Basic High Availability Concepts," presents the lan(cid:173)
`guage used to describe highly available systems and components and
`introduces ways of measming availability.
`
`Chapter 2, "Creating a High Availability Cluster," describes in more
`detail the principles of HA configuration, with examples.
`
`Chapter 3, "HP's High Availability Cluster Components," is an over(cid:173)
`view of HP's current roster of high availability software and hard(cid:173)
`ware offerings.
`
`• Chapter 4, "Sample HA Solutions," discusses afew concrete exam(cid:173)
`ples of highly available cluster solutions.
`
`Chapter 5, "Glossary," gives definitions of important words and
`phrases used to describe high availability.
`
`Additional information is available in the HP publications Managing
`MC!ServiceGuard and Configuring OPS Clusters with MC!LockMan(cid:173)
`ager. The HP 9000 Servers Configuration Guide contains detailed
`information about supported high availability configurations. This and
`other more specialized documents on enterprise clusters are available
`from your HP representative.
`
`XV
`
`DHPN-1003 I Page 15 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`- - - - -------
`Acknowledgments
`
`This book has benefited from the careful review of many individuals
`inside and outside of Hewlett-Packard. The author gratefully acknowl(cid:173)
`edges the contributions of these colleagues, many of whom are listed
`here: Joe Algieri, Sally Anderson, Joe Bac, Bob Baird, Trent Bass, Dan
`Beringer, Claude Brazell, Thomas Buenermann, Xuan Bui, Karl-Heinz
`Busse, Bruce Campbell, Larry Cargnoni, Gina Cassinelli, Marian
`Cochran, Annie Cooperman, Ron Czinski, Dan Dickerman, Pam Dick(cid:173)
`erman, Larry Dino, Janie Felix, John Foxcroft, Shivaji Ganesh, Janet
`Gee, Mike Gutter, Terry Hand, Michael Hayward, Frank Ho, Margaret
`Hunter, Lisa Iarkowski, Art Ipri, Michael Kahn, Marty King, Clark
`Macaulay, Gary Marcos, Debby Mcisaac, Doug McKenzie, Tim Met(cid:173)
`calf, Parissa Mohamadi, Alex Morgan, Markus Ostrowicki, Bob Ramer,
`Bob Sauers, Wesley Sawyer, David Scott, Dan Shive, Christine Smith,
`Eric Soderberg, Steve Stichler, Tim Stockwell, Brad Stone, Liz Tam,
`Bob Togasaki, Emil Velez, Tad Walsh, and Bev Woods. A special thank
`you goes to those groups of Hewlett-Packard customers who read and
`commented on early versions of the manuscript. Enors and omissions
`are the author's sole responsibility.
`
`xvi
`
`DHPN-1003 I Page 16 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`About the Author
`
`About the Author
`
`PeterS. Weygant is a Leaming Products Engineer in the General Sys(cid:173)
`tems Solutions laboratory at Hewlett-Packard. Formerly a professor of
`English, he has been a technical writer and consultant in the computer
`industry for the last 15 years. He has developed documentation and
`managed publication projects in the areas of digital imaging, relational
`database technology, and high availability systems. He has a BA degree
`in English Literature from Colby College as well as MA and PhD
`degrees in English from the University of Pennsylvania.
`
`xvii
`
`DHPN-1003 I Page 17 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`CHAPTER 1
`Basic High Availability
`Concepts
`
`This book takes an elementary look at high availabil(cid:173)
`ity (HA) computing and how it is implemented through
`enterprise-level cluster solutions. We start in this chapter
`with some of the basic concepts of HA. Here's what we'll
`cover:
`
`• What is High Availability?
`• High Availability as a Business Requirement
`• What Are the Measures of High Availability?
`• Understanding the Obstacles to High Availability
`• Preparing Your Organization for High Availability
`• The Starting Point for a High Availability System
`• From High Reliability to High Availability
`• Designing a Highly Available System
`
`1
`
`DHPN-1003 I Page 18 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Basic High Availability Concepts
`
`Later chapters explore the implementation of high
`availability in clusters, then describe HP's high availability
`products in more detail. A separate chapter is devoted to
`concrete examples of business solutions that use HA.
`
`What is High Availability?
`
`Before exploring the implications of high availability
`in computer systems, we need to define some terms. What
`do we mean by phrases like "availability," "high availabil(cid:173)
`ity," and "high availability computing?"
`
`Available
`
`The term available describes a system that provides a
`specific level of service as needed. This idea of availability
`is part of everyday thinking. In computing, availability is
`generally understood as the period of time when services
`are available (for instance, 16 hours a day, six days a week)
`or as the time required for the system to respond to users
`(for example, under 1 second response time). Any loss of
`service, whether planned or unplanned, is known as an
`outage. Downtime is the duration of an outage measured
`in units of time (e.g., minutes or hours).
`
`2
`
`DHPN-1003 I Page 19 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`What is High Availability?
`
`Highly Available
`
`Figure 1. 1 Highly Available Services: Electricity
`
`Highly available characterizes a system that is
`designed to avoid the loss of service by reducing or manag(cid:173)
`ing failures as well as minimizing planned downtime for
`the system. We expect a service to be highly available when
`life, health, and well-being, including the economic well(cid:173)
`being of a company, depend on it.
`
`3
`
`DHPN-1003 I Page 20 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Basic High Availability Concepts
`
`For example, we expect electrical service to be highly
`available. All but the smallest, shortest outages are unac(cid:173)
`ceptable, since we have geared our lives to depend on elec(cid:173)
`tricity for refrigeration, heating, and lighting, in addition to
`less important daily needs.
`
`Even the most highly available services occasionally
`go out, as anyone who has experienced a blackout or
`brownout in a large city can attest. But in these cases, we
`expect to see an effort to restore service at once. When a
`failure occurs, we expect the electric company to be on the
`road fixing the problem as soon as possible.
`
`Highly Available Computing
`
`In many businesses, the availability of computers has
`become just as important as the availability of electric
`power itself. Highly available computing uses computer
`systems which are designed and managed to operate with only
`a small amount of planned and unplanned downtime.
`
`Note that highly available is not an absolute. The needs
`of different businesses for high availability are quite
`diverse. International businesses or companies running
`multiple shifts may require user access to databases around
`the clock. Financial institutions must be able to transfer
`funds at any time of night or day, seven days a week. On
`the other hand, some retail businesses may require the
`
`4
`
`DHPN-1003 I Page 21 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`What is High Availability?
`
`Figure 1.2 Service Outage
`
`computer to be available only 18 hours a day, but during
`these 18 hours they may require sub-second response time
`for transaction processing.
`
`Service Levels
`
`The service level of a system is the degree of service
`the system will provide to its users. Often, the service level
`is spelled out in a document known as a service level agree(cid:173)
`ment (SLA). The service levels your business requires
`determine the kind of applications you develop, and high
`availability systems provide the hardware and software
`
`5
`
`DHPN-1003 I Page 22 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Basic High Availability Concepts
`
`framework in which these applications can work effec(cid:173)
`tively to provide the needed level of service. High avail(cid:173)
`ability implies a service level in which both planned and
`unplanned computer outages do not exceed a small stated
`value.
`
`Continuous Availability
`Continuous availability means non-stop service, that
`is, there are no planned or unplanned outages at all. This is
`a much more ambitious goal than high availability, since
`there can be no lapse in service. In effect, continuous avail(cid:173)
`ability is an ideal state rather than a characteristic of any
`real world system.
`
`The term is sometimes used to indicate a very high
`level of availability in which only a very small known
`quantity of downtime is acceptable. Note that high avail(cid:173)
`ability does not imply continuous availability.
`
`Fault Tolerance
`Fault tolerance is not a degree of availability so much
`as a method for achieving very high levels of availability. A
`fault tolerant system is characterized by redundancy in
`most of the hardware components, including CPU, mem(cid:173)
`ory, I/0 subsystems, and other elements. A fault tolerant
`system is one that has the ability to continue service in spite
`of a hardware or software failure. However, even fault tol(cid:173)
`erant systems are subject to outages from human error.
`Note that high availability does not imply fault tolerance.
`
`6
`
`DHPN-1003 I Page 23 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`What is High Availability?
`
`Matching Availability to User Needs
`
`A failure affects availability when it results in an
`unplanned loss of service that lasts long enough to create a
`problem for users of the system. User sensitivity will
`depend on the specific application. For example, a failure
`that is corrected within one second may not result in any
`perceptible loss of service in an environment that does on(cid:173)
`line transaction processing (OLTP); but for a scientific
`application that runs in a real-time environment, one sec(cid:173)
`ond may be an unacceptable interval.
`
`Since any component can fail, the challenge is to
`design systems in which problems can be predicted and
`isolated before a failure occurs and in which failures are
`quickly detected and corrected when they happen.
`
`Choosing a Solution
`
`Your exact requirements for availability determine the
`kind of solution you need. For example, if the loss of a sys(cid:173)
`tem for a few hours of planned downtime is acceptable to
`you, then you may not need to purchase storage products
`with hot pluggable disks. On the other hand, if you cannot
`afford a planned period of maintenance during which a
`disk replacement could be done on a mirrored disk system,
`then you may wish to consider a HA disk array that sup(cid:173)
`ports hot plugging or hot swapping of components.
`(Descriptions of these HA products appear in later sec(cid:173)
`tions.)
`
`7
`
`DHPN-1003 I Page 24 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Basic High Availability Concepts
`
`High Availability as a Business
`Requirement
`
`In the current business climate, high availability com(cid:173)
`puting is often seen as a requirement, not a luxury. From
`one perspective, high availability is a form of insurance
`against the loss of business due to computer downtime.
`From another point of view, high availability provides new
`opportunities by allowing your company to provide better
`and more competitive customer service.
`
`High Availability as Insurance
`
`High availability computing is often seen as insurance
`against the following kinds of damage:
`
`• Loss of income
`• Customer dissatisfaction
`• Missed opportunities
`
`For commercial computing, a highly available solu(cid:173)
`tion is needed when loss of the system results in loss of rev(cid:173)
`enue. In such cases, the application is said to be mission(cid:173)
`critical. For all mission-critical applications- that is, where
`income may be lost through downtime - high availability
`is a requirement. In banking, for example, the ability to
`obtain certain account balances 24 hours a day may be mis(cid:173)
`sion-critical. In parts of the securities business, the need for
`
`8
`
`DHPN-1003 I Page 25 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`High Availability as a Business Requirement
`
`high availability may only be for that portion of the day
`when the stock market is active; at other times, systems
`may be safely brought down.
`
`High Availability as Opportunity
`
`Highly available computing provides a business
`opportunity, since there is an increasing demand for
`"around the clock" computerized services in areas as
`diverse as banking and financial market operations, com(cid:173)
`munications, order entry and catalog services, resource
`management, and others. It is not possible to give a simple
`definition of when an application is mission-critical or of
`when high availability of the application creates new
`opportunities; this depends on the nature of the business.
`However, in any business that depends on computers, the
`following principles are always true:
`
`• The degree of availability required is determined by
`business needs. There is no absolute amount of
`availability that is right for all businesses.
`
`• There are many ways to achieve high availability.
`
`• The means of achieving high availability affects all
`aspects of the system.
`
`• The likelihood of failure can be reduced by creating
`an infrastructure that stresses clear procedures and
`preventive maintenance.
`
`• Recovery from failures must be planned.
`
`9
`
`DHPN-1003 I Page 26 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Basic High Availability Concepts
`
`Some or all of the following are expectations for the
`software applications that run in mission-critical environ(cid:173)
`ments:
`
`• There should be a low rate of application failures,
`that is, a maximum time between failures.
`
`• Applications should be able to recover after failure.
`
`• There should be minimal scheduled downtime.
`
`• The system should be configurable without shut(cid:173)
`down.
`
`• System management tools must be available.
`
`Cost of High Availability
`
`As with other kinds of insurance, the cost depends on
`the degree of availability you choose. Thus the value of
`high availability to the enterprise is directly related to the
`costs of outages. The higher the cost of outage, the easier it
`becomes to justify the expense of high availability solu(cid:173)
`tions. As the degree of availability approaches the ideal of
`100% availability, the cost of the solution increases more
`rapidly. Thus, the cost of 99.95% availability is significantly
`greater than the cost of 99.5% availability, and the cost of
`99.5% availability is significantly greater than 99% avail(cid:173)
`ability, and so on.
`
`10
`
`DHPN-1003 I Page 27 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`What Are the Measures of High Availability?
`
`What Are the Measures of High
`Availability?
`
`Availability and reliability can be described in
`terms of numbers, though doing so can be very mis(cid:173)
`leading. In fact, there is no standard method for model(cid:173)
`ing or calculating the degree of availability in a
`computer system. The important thing is to create clear
`definitions of what the numbers mean and then use
`them consistently. Remember that availability is not a
`measurable attribute of a system like CPU clock speed.
`Availability can only be measured historically, based on
`the behavior of the actual system. Moreover, in measur(cid:173)
`ing availability, it is important to ask not simply, "Is the
`application available?" but "Is the entire system pro(cid:173)
`viding service at the proper level?"
`
`Availability is related to reliability, but they are not the
`same thing. Availability is the percentage of total system
`time the computer system is accessible for normal usage.
`Reliability is the amount of time before a system is
`expected to fail. Availability includes reliability.
`
`Calculating Availability
`
`The formula in Figure 1.3 defines availability as the
`percentage of elapsed time that a unit can be used. Elapsed
`time is continuous time (operating time+ downtime).
`
`11
`
`DHPN-1003 I Page 28 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Basic High Availability Concepts
`
`% Availabilit =
`y
`
`(Total Elapsed Time -Sum
`of Inoperative Times)
`------------------
`Total Elapsed Time
`
`Figure 1.3 Availability
`
`Availability is actually the probability that a unit is
`available (that is, operating normally). Availability is usu(cid:173)
`ally expressed as a percentage of hours per w eek, month, or
`year during which the system and its services can be used
`for normal business.
`
`Expected Period of Operation
`
`Measures of availability must be seen against the
`background of the organization's expected period of opera(cid:173)
`tion of the system. The following tables show the actual
`
`12
`
`DHPN-1 003 I Page 29 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`What Are the Measures of High Availability?
`
`hours of uptime and downtime associated with different
`percentages of availability for two common periods of
`operation. Table 1.1 shows 24x7x365, which stands for a
`system that is expected to be in use 24 hours a day, seven
`days a week, 365 days a year.
`
`Table 1. 1 Uptime and Downtime for a 24x7x365 System
`
`Availability Minimum Maximum
`Allowable
`Expected
`Downtime
`Uptime
`
`Remaining
`Time
`
`99%
`
`99.5%
`
`99.95%
`
`100%
`
`8672
`
`8716
`
`8755
`
`8760
`
`88
`
`44
`
`5
`
`0
`
`0
`
`0
`
`0
`
`0
`
`This table shows that there is no remaining time on the
`system at all. All the available time in the year (8760 hours)
`is accounted for. This means that all maintenance must be
`carried out either when the system is up or during the
`allowable downtime hours. In addition, the higher the per(cid:173)
`centage of availability, the less time is allowable for failure.
`
`Table 1.2 shows a 12x5x52 system, which is expected
`to be up for 12 hours a day, five days a week, 52 weeks a
`year.
`
`13
`
`DHPN-1003 I Page 30 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`Basic High Availability Concepts
`
`Table 1.2 Uptime and Downtime for a 12x5x52 System
`
`Availability Minimum
`Expected
`Uptime
`
`Maximum
`Allowable
`Downtime
`
`Remaining
`Time
`
`99%
`
`99.5%
`
`99.95%
`
`100%
`
`3088
`
`3104
`
`3118
`
`3118
`
`32
`
`16
`
`2
`
`0
`
`5642
`
`5642
`
`5642
`
`5642
`
`This table shows that for the 12x5x52 system, there are
`5642 hours of remaining time, which can be used for
`planned maintenance operations requiring the system to be
`down.
`
`Calculating Mean Time Between Failures
`
`Availability is related to failure rates of system compo(cid:173)
`nents. A common measure of equipment reliability is the
`mean time between failures (MTBF). This measure is usu(cid:173)
`ally provided for individual system components, such as
`disks. Measures like these are useful, but they are only one
`dimension of the complete picture of high availability. For
`example, they do not take into account the differences in
`recovery times after failure.
`
`MTBF is given by the formula shown in Figure 1.4.
`
`14
`
`DHPN-1003 I Page 31 of 181
`
`IPR2014-00901 Owner Ex. 2101
`ETRI, Patent Owner
`VMware, Petitioner
`
`
`
`What Are the Measures of High Availability?
`
`MTBF=
`
`Total Operating Time
`
`Total No. of Failures
`
`Figure 1.4 Mean Time Between Failures
`
`The MTBF is calculated by summing the actual operat(cid:173)
`ing times of all units, including units that do not fail, and
`dividing that sum by the sum of all failures of the units.
`Operating time is the sum of the hours when the system is
`in use (that is, not
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
