`
`-IIIIIIIIIII I
`
`/
`
`LEGENT
`
`ESMTM : Product Introduction
`
`Document Number ES-LAN22—1
`
`Release 1.0 for 08/2
`
`This manual provides an
`introduction to Enterprise
`Storage Manager.
`
`Trademark of Legent Software, Inc.
`
`EMCVMW 1026
`
`
`
`The information in this document applies to Release 1.0 of the ESM program product for the 05/2 operating system.
`
`Copyright © 1994 Mission Critical Software, Inc., as an unpublished work. This notice does not imply unrestricted or
`public access to these materials which are trade secret of Legent Corporation or its subsidiaries or affiliates (together
`referred to as "Legent"), and which may not be reproduced, used, sold or transferred to any third party without
`LEGENT's prior written consent.
`
`All rights reserved.
`
`RESTRICTED RIGHTS LEGEND
`
`Use, duplication, or disclosure by the Government is subject to restrictions as set forth in subparagraph (c) (1) (ii) of the
`Rights in Technical Data and Computer Software clause at DFARS 252.227—7013.
`
`First Edition - April 1994 (ES—LAN22-1)
`
`(ESM Release 1.0)
`
`Published by LEGENT Corporation
`2000 Park Lane, Pittsburgh, Pennsylvania 15275
`Phone: (412) 924—2500
`
`Technical Support: (713) 974-1500 x 130 Fax: (713) 975-7565
`
`©1993 Mission Critical Software, Inc.
`
`
`
`
`
`About this manual
`
`This manual provides an overview of Legent's Enterprise Storage Manager, including:
`
`0
`
`0
`
`0
`
`0
`
`0
`
`o
`
`o
`
`Introduction
`
`Benefits
`
`Features
`
`Product Description
`
`Comparison of ESM to Traditional Approaches
`
`Platforms Supported by ESM
`
`ESM Installation
`
`
`
`
`
`Related Publications
`
`ESM/LAN Technical Overview
`
`ESM/LAN Installation and Customization
`
`ESM/LAN Administrator's Guide
`
`ESM/LAN Administrator's Reference
`
`ESM/LAN User's Guide
`
`ESM/LAN Messages and Codes
`
`ESM/CSP for MVS Overview
`
`ESM/CSP for MVS Installation and Customization
`
`ESM/CSP for MVS Administrator's Guide
`
`ESM/CSP for MVS Administrator's Reference
`
`ESM/CSP for MVS Messages and Codes
`
`iv
`
`
`
`
`
`Table of Contents
`
`Introduction
`
`The Need for Better Tools for LAN Storage Management ..................
`Enterprise Storage Management Requirements .................................
`The ESM Solution ........................................................................
`
`ESM Automates LAN Storage Management .....................................
`Who Should Use ESM? .................................................................
`
`ESM Benefits
`
`Automated Protection Against Loss of Critical Data ..........................
`Rapid Recovery from Data Loss .....................................................
`Protection Against Data Disaster ....................................................
`Savings in Personnel Costs ...........................................................
`Cost-Effective Use of Resources ....................................................
`
`Better User Service and Enhanced User Productivity .........................
`Unlimited File Backup ...................................................................
`Enforcement of Standards .............................................................
`
`ESM Features
`
`Key Features ...............................................................................
`
`Double Backup
`Advantages ................................................................................
`Full and incremental backups ........................................................
`Primary on hard disk ....................................................................
`Minimal space, maximum benefits ..................................................
`Mainframe security ......................................................................
`
`Duplicate File Elimination
`Binary Objects (BOBs) ..................................................................
`The BOBlD ..................................................................................
`BOB benefits ...............................................................................
`
`Audits - Ensuring the Recoverability of Lost Data
`Automated Local Server Rebuild
`
`Management Classes
`
`Backup Retention Patterns
`
`Priority-based Backup
`
`Data Compression
`
`Selective Backup
`Co-location of Restorable Files
`
`Persistence and Restartability
`
`Comprehensive Restart and Recovery
`
`Security
`
`Automated System Operator (ASO)
`
`Reliability, Availability and Serviceability (RAS)
`
`
`
`_ Reporting
`
`Product Description
`
`Multiple Server Sites
`
`Single Server Sites
`
`3-10
`
`4—1
`
`4-1
`
`4-2
`
`4-3
`ESM Components
`LAN Subsystem ........................................................................... 4-3
`Mainframe Subsystem .................................................................. 4-4
`
`Backup of LAN Server Data
`
`5-1
`
`Verification of Backup Data Integrity .............................................. 5-1
`Transferring Backup Data Off-Site .................................................. 5-2
`Recovery from Individual LAN Server Failures .................................. 5-2
`Site Disaster Recovery ................................................................. 5—2
`Retrieval of Prior Versions of Files .................................................. 5-3
`
`Platforms Supported by ESM
`
`ESM Installation
`
`6-1
`
`7-1
`
`ESM Server ................................................................................. 7—1
`ESM/GU! .................................................................................... 7-1
`ESM/CSP for MVS ....................................................................... 7-1
`
`vi
`
`
`
`
`
`Figures
`
`1-1. Growth of WAN storage ................................................................... 1—2
`2-1. Dual Backup Process ........................................................................ 2—1
`3—1. BOB Processing ............................................................................... 3-4
`3—2. MGMT.REXX .................................................................................. 3-6
`
`3—3. Management Summary Report ........................................................... 3—10
`4-1. ESM Multiple-Server Site Configuration ............................................... 4—1
`4—1 ESM Single—Server Site Configuration .................................................. 4-2
`4-3 ESM Components ............................................................................. 4-3
`4—4 ESM CSP ......................................................................................... 4-4
`
`vii
`
`
`
`
`
`Trademark Information
`
`This manual refers to the following manufacturers' products:
`
`Digital Communications Associates, Inc.
`DCA®
`DCA Communication Server
`DCA Communication Workstation
`
`International Business Machines Corporation
`ACF/VTAMTM
`
`Advanced Peer—to-Peer NetworkingTM
`AnyNetTM
`Communications Manager/2
`DFSMS/MVSTM
`ES/3090TM
`ES/9000TM
`Extended Services for 08/2
`LAN Server
`MVS/ESATM
`NetView®
`IBM®
`
`Operating System/2®
`OS/2®
`OS/2 Peer Services
`OS/2 Extended Edition
`S/390®
`VTAM®
`
`Intel Corporation
`80386 microprocessor
`80486 microprocessor
`
`Legent® Corporation
`Legent®
`
`Microsoft® Corporation
`WindowsTM
`Windows NT
`
`LAN Manager
`
`SAS Institute Inc.
`SAS®
`
`Systems Center, Inc.
`NETMASTER®
`
`viii
`
`
`
`The Need for Better
`
`Tools for LAN Storage
`Management
`
`
`
`Introduction
`
`A decade ago, the computer data critical to the survival and success of an enterprise resided
`on high-reliability disk drives in a secure corporate data center, staffed around the clock by
`skilled systems management specialists. Today, mission—critical corporate data can be found
`on everything from LAN servers in branch offices to workstations in employee homes to
`laptops rattling around the back seats ofjeeps in the desert. The volume of distributed storage
`in use is exploding as mass production gears up and unit costs come down. Total disk space
`in Intel-based PCs has grown exponentially from virtual insignificance in the early 19803 to
`an estimated 7 billion megabytes in 1993, increasing to 37 billion megabytes by 1997. As
`enterprises rely increasingly on distributed data and disk space for vital operations, these
`assets must be managed.
`
`With the price and performance of microprocessor-based computers outpacing that of
`mainframes, large organizations are adding much of their new computing capacity in the form
`of networked PCs and workstations. These organizations are beginning to appreciate the
`systems management problems and high support and user costs associated with these new
`platforms, such as:
`'
`
`X
`
`PC-type tape drives for LAN backup have a number of drawbacks. They are:
`
`® Much less reliable than mainframe tape drives because they are manufactured to
`a much lower price point.
`
`® Require attended operations by office personnel.
`
`® Typically very slow.
`
`® Require frequent cleaning to maintain serviceability - care that is difficult to
`assure in an office environment.
`
`X
`
`X
`
`No way exists to verify backups are made and tapes are taken offsite for disaster
`recovery.
`
`Recent "life cycle" cost analyses suggest hidden costs, such as time spent by users
`performing administrative functions, may make the total cost of distributed
`computing much higher than originally estimated. Studies indicate user time spent
`on systems management is a major reason that the massive investment in desktop
`computing over the past 10 years has produced little measurable improvement in
`corporate productivity.
`
`Chapter 1.
`
`Introduction
`
`1-1
`
`
`
`Enterprise Storage
`Management
`Requirements
`
`Data and disk space management, including backup and restore, is the single most critical
`computer system management task. Large organizations need an enterprise solution for
`managing distributed storage meeting the following requirements:
`
`0
`
`Rapid restore capability
`
`Avoids long disruption of business operations if
`active data is destroyed or corrupted. This is
`especially critical
`for LAN servers dispersed
`over hundreds or thousands of remote locations,
`with
`littie
`skilied administrative
`assistance
`
`locally.
`
`0 Automatic auditing
`
`Ensures backup data is restorable.
`
`0 Automatic vaulting
`
`Moving backup data to a secure central location
`in case of local site disaster.
`
`0
`
`Performance capabilities
`
`Addressing limited wide area network (WAN)
`bandwidth during backup and restore, such as
`automatic duplicate data elimination, backing up
`only changed data,
`and data
`compression
`techniques.
`
`Centralized reporting
`
`Verifies all data is protected.
`
`Deals with changing configurations, operational
`problems, and reliability issues
`inherent
`in
`heterogeneous computing networks; protects
`against human error; and controls the growth and
`cost of skilled technical staff.
`
`Manages enterprises with hundreds of thousands
`of hard disks, millions of directories, and billions
`of files.
`
`and recovery
`significant backup
`Transfers
`processing to previously under-utilized network
`resources and exploits the differences in MIPS
`and
`disk
`space
`costs
`between PCs
`and
`mainframes.
`
`o
`
`0
`
`Built-in automated systems
`
`0
`
`Scalable architecture
`
`0 Advanced design
`
`1—2
`
`ESM Product Introduction
`
`
`
`The ESM Solution
`
`
`
`Legent Corporation's Enterprise Storage Manager (ESM) meets the requirements of enterprise
`storage management.
`Incorporating unique, patent-pending technology developed
`specifically for this purpose, ESM manages LAN servers connected to mainframe data
`centers through Wide Area Networks (WANs).
`
`The use of WAN versus Mainframe storage has increased significantly in the last few years
`and promises continued growth.
`
`WAN Storage Growth
`
`80Percent Growth
`
`
`
`+ Mainframe WAN
`
`1996
`
`1 993
`
`1994
`
`1995
`
`Figure 1-1. Growth of WAN storage.
`
`ESM's unequaled approach to eliminating duplicate data dramatically reduces LAN disk
`storage and bandwidth requirements for transmission to mainframe storage. ESM automates
`and assures the performance of a number of critical LAN system management functions,
`including data backup and movement to offsite storage, and recovery from server failures.
`
`ESM requires no dedicated support personnel at the LAN site for normal operation and does
`not use PC—type tape drives or other specialized hardware. ESM manages OS/Z-based LAN
`servers, and provides support for other servers through the Universal Naming Convention
`(UNC) protocol that allows a great degree of file visibility over the network.
`
`ESM Automates LAN
`
`Storage Management
`
`ESM is a Won, reducing the risks associated with large-scale LAN use by
`building on the strengths of the existing mainframe data center.
`It provides dependable,
`industrial-strength data backup and restore for LANs, without the need for on-site expertise or
`dedicated personnel to perform and monitor these activities.
`
`ESM:
`
`s/ Takes responsibility for the most critical LAN systems management tasks, backup
`and offsite vaulting of data, and recovery from site disasters, hardware failures, and
`user errors.
`
`s/ Performs backup tasks automatically.
`
`0/ Provides capabilities to quickly restore business operations interrupted by data
`disaster.
`
`Chapter 1.
`
`Introduction
`
`1'3
`
`
`
`V ls unique in its ability to minimize or eliminate the costs of prime-shift processing by
`performing operations during non-prime periods,
`lowering incremental costs of
`mainframe, telecommunications, and LAN resource usage.
`
`1/
`
`Incorporates a number of unique approaches to minimize telecommunications
`bandwidth requirements.
`
`reducing costs even when using computing and
`V Saves expensive staff time,
`telecommunications resources during prime shift.
`
`ESM looks at the storage management problem from the point of view of systems support
`professionals, who are responsible for making sure that systems management actually
`happens and that the enterprise‘s data is protected.
`
`Who Should Use ESM?
`
`For administrators managing emerging and large-scale LANs and distributed computing
`networks, ESM is the ideal solution for automating backup and restore tasks for LANs
`connected to traditional mainframe data centers over leased or dial-up telecommunication
`links. ESM operates effectively on a single-server LAN, and is optimized for installations
`with two or more servers.
`
`.5,”V
`
`.,
`
`1-4
`
`ESM Product Introduction
`
`
`
`
`
`ESM Benefits
`
`Enterprise Storage Manager is a full-featured, robust, distributed application that benefits
`your entire enterprise. ESM automatically protects distributed enterprise information from
`loss or destruction at minimum personnel and resource cost, minimizes the time to recover
`from data loss situations, increases user productivity, enhances service to users, and helps
`enforce data management standards throughout the enterprise.
`It accomplishes these tasks
`with few resources and can manage virtually unlimited amounts of data.
`
`
`Automated Protection
`Against Loss of Critical
`Data
`
`ESM protects data by completely automating the backup process, by saving the data in two
`locations, on the LAN and at the corporate data center, and by automatically auditing the data
`as it is saved.
`
`ESM automation facilities initiate the backup process without manual intervention. Daily
`backups are scheduled for most files, and checkpoint backups for critical, time-dependent
`files may be scheduled automatically.
`
`A dual backup process ensures data availability.
`
`SERVER l
`
`SERVER 2
`
`Figure 2-1. Dual Backup Process.
`
`Chapter 2. ESM Benefits
`
`2-1
`
`
`
`The first backup is to LAN disk storage. ESM ensures restorability by performing random
`audits of the data as it is backed up, comparing backup data to the original file.
`
`For added security, ESM transfers a copy of the data to the corporate data center, which is
`unmatched in its ability to assimilate large volumes of data over telecommunications lines,
`process and verify that data, and store it on tape for later retrieval. With high-speed, high-
`reliability automated tape systems, mainframes can locate and retrieve data files from tape
`much faster than LANs, and are physically secure in ways that are difficult, if not impossible,
`to duplicate in typical office environments. Corporate data centers also provide established
`procedures for off-site disaster storage, a critical element in the ultimate protection of mission
`critical information.
`
`In contrast, backups can be missed due to employee absence or inattention in the LAN
`environment, or backup tapes might not be moved off site to guard against a physical disaster.
`Disgruntled employees can easily cause the irretrievable loss of key data. The tape cartridges
`used for backup are so small and easy to conceal, that they (and your data) are easy prey for
`thieves and industrial spies.
`
`In addition, the PC tape drives currently used for LAN backup are simply not reliable enough
`to safeguard critical data. PC tape drives also require frequent cleaning and maintenance,
`attention that they may not receive in a typical office environment. ESM, however, requires
`no PC—type tape drives, eliminating exposure to data loss.
`’
`
`When a mission critical application is interrupted due to data loss, it is imperative that it be
`brought back on-line as quickly as possible by restoring corrupted or deleted files. ESM
`minimizes downtime by storing backup data on LAN disk storage for rapid restore by users or
`the LAN administrator. With local backup, most data can be restored in seconds. Both
`graphical and command line interfaces are provided to allow users to initiate the rapid restore
`process.
`
`In the event that the data is not available on the LAN, ESM automatically recognizes the need
`to obtain the data from the corporate data center, sends it to the LAN, and restores it without
`manual intervention or delays.
`
`Computer professionals accept the inevitability that data will be lost. Hardware components
`fail and software components have bugs. People mistakenly erase files, directories, or even
`drives.
`In extreme but all too familiar cases, natural disasters or acts of sabotage result in the
`loss of a complete server disk, a complete server or an entire site.
`
`ESM provides protection against disasters by vaulting data at the mainframe data center, and
`by providing a specialized disaster recovery facility, server rebuild. Server rebuild can be
`used to restore a server to its previous state, or to reconstruct the server software on a new
`machine. You can rebuild entire sites by reconstructing each server from ESM backup data.
`
`Rapid Recovery from
`Data Loss
`
`Protection Against
`Data Disaster
`
`Savings in Personnel
`Costs
`
`Today, LAN administrators are responsible for backup and recovery of the distributed
`environment, a cost that can be staggering. Increases in support personnel costs for distributed
`computing administration are the leading cause of increases in the cost of distributed
`computing resulting from downsizing of host-based applications. Controlling these costs is
`difficult, because they often are not measured or welluunderstood.
`
`2-2
`
`ESM Product introduction
`
`
`
`support personnel costs
`reduces
`ESM, by automating LAN backup and recovery,
`significantly.
`Instead of using administrators at each LAN or distributed site to perform file
`backup and recovery operations, ESM automates the backup process, and permits users to
`initiate restore processes as needed. While administrators may be needed to assist in disaster
`situations when server disks or entire servers are destroyed, ESM takes the critical first step
`toward the extended data center, which will permit distributed computing with no on-site
`support personnel or expertise.
`
`Cost-Effective Use of
`Resources
`
`ESM features minimize the use of resources on both the LAN and mainframe host system,
`providing the most cost-effective method available today for safe, secure disaster backup and
`recovery for distributed environments. Important cost-saving features include:
`
`1/ Use of the LAN disk for primary backup, at a cost of less than one-fifth that of
`mainframe disk storage. Additional savings in mainframe processing and network
`usage are realized during restore operations whenever LAN-based backup copies are
`used.
`
`V An approach to eliminating duplicate copies of data files that dramatically reduces
`the need for LAN disk storage and transmission bandwidth between the LAN and
`host mainframe.1 This approach relies not on file names, sizes, or creation dates, but
`rather on a sophisticated technique that identifies unique file segments based on file
`content. There is no equal to this approach in the industry today!
`
`3/ Unique data compression algorithms that further reduce the need for LAN disk
`storage and LAN to mainframe transmission bandwidth.
`
`\ Minimizing the use of mainframe processing and storage resources.
`
`‘ The use of tape storage for backup and archive on the mainframe, rather than
`expensive disk storage. At an estimated $.01 per megabyte for tape storage today,
`the savings associated with this approach are significant.
`
`V LAN implementation that allows a single communications server per LAN to be
`used for data transfer with the mainframe. This results in reduced overhead and
`
`savings of communications hardware and software costs that would be required if
`each server required such a connection.
`
`You can run ESM off-prime-shifi, when both LAN and mainframe resource requirements are
`at their minimum. With the capability to automatically prioritize the order of data backup, so
`that the most critical information is saved first, and with ESM's ability to interrupt processing
`whenever the backup window has expired and resume processing at the next off-shift
`window, ESM stands alone in its ability to minimize or eliminate the costs of prime—shift
`processing.
`
`Better User Service
`and Enhanced User
`
`Productivity
`
`Studies show that 80 percent of data loss situations result from user error, such as pressing the
`wrong key or performing the wrong function or the right function on the wrong data.
`
`ESM enhances user productivity by automatically backing up and archiving user files, and
`allows users to initiate backup or archive operations as needed. When a data loss situation
`occurs, users can recover by browsing backup copies of their files and restoring available
`versions without the inherent delays that result from interaction with overloaded help desk
`personnel. Users get faster service and enhanced productivity.
`
`1US. Patents Applied for.
`
`Chapter 2. ESM Benefits
`
`2-3
`
`
`
`Unlimited File Backup
`
`Enforcement of
`Standards
`
`is
`Scalability, a critical requirement for managing today’s mission critical applications,
`guaranteed with ESM. As more servers and storage are added, ESM's capability to manage
`them grows in step. Also, because ESM is a true cooperative processing application, ESM
`LAN processes may be dispatched to multiple LAN servers to distribute the load more evenly
`or to take advantage of faster processors or enhanced disk storage capabilities.
`
`Enterprise-wide duplicate elimination and compression also means, optimal use of disk
`storage and expensive (and relatively slow) transmission facilities between the LAN and the
`mainframe site.
`
`ESM supports a standard methodology for handling different types of files by implementing
`management classes. Enterprises with heavy investments in Systems Managed Storage
`(SMS) over the last several years will find the concept and implementation of ESM’s
`management class facility familiar. Through this facility, priorities may be assigned to
`different types of files, causing ESM to back up and restore the highest-priority files first.
`In
`addition, this facility permits the designation of specific retention periods for some files,
`while ignoring other files in the backup process altogether. ESM provides an almost
`unlimited capability to develop and enforce standards for file backup.
`
`2-4-
`
`ESM Product Introduction
`
`
`
`
`
`ESM Features
`
`ESM is based on a unique LAN-centric design that allows a LAN to manage its own data
`while providing the necessary off-site backup at your corporate data center. The entire
`process is centrally administered through automated data management policies that may be
`changed locally if authorized by central
`site administration.
`These automated data
`management policies control LAN storage management processing, including:
`
`. © Policy implementation.
`
`© Recovery from exception conditions.
`
`© Automation of routine tasks.
`
`Key Features
`
`ESM is a genuine cooperative processing application that makes effective use of all available
`computing resources to manage the disk storage on your LAN servers. For example, ESM
`often uses the CPU power of several servers to compress the various file segments that
`comprise one large file. By balancing load across available CPUs, ESM avoids bottlenecks
`that might be encountered with CPU-intensive operations such as compression.
`
`Key features of ESM include:
`
`Double Backup
`
`\“\\\‘\‘\“\\\\
`
`Duplicate File Elimination
`
`Audits - Ensuring the Recoverability of Lost Data
`Automated Local Server Rebuild
`
`Management Classes
`
`Backup Retention Patterns
`
`Priority-based Backup
`
`Data Compression
`
`Selective Backup
`Co-location of Restorable Files
`
`Persistence/Restartability
`
`Comprehensive Restart/Recovery
`
`Security
`
`Automated System Operator (ASO)
`
`Reliability/Availability/Serviceability (RAS)
`
`Reporting.
`
`Chapter 3. ESM Features 3-1
`
`
`
`
`
`Double Backup
`
`In addition to
`ESM backs up all LAN server data to two separate locations at the same time.
`an incremental backup to the mainframe data center, ESM backs up the data to local disk
`storage on the other servers on the LAN, thus creating "local compressed backup."
`
`Local compressed backup gives ESM many unique functional advantages. Certain critical
`files can be backed up more than once a day (checkpoint backup), minimizing lost data in the
`event of a server failure. When one disk on the LAN is lost, a replacement disk is rebuilt and
`restored from local compressed backup faster than any restore-from-tape approach and is not
`a function of the bandwidth between the mainframe data center and the LAN site. Also, 80
`percent of user requests for restores of individual files are for the most recent backup copy.
`With ESM, this data will usually be available locally, providing faster fulfillment of the user
`request without the need to restore the backup file over a telecommunications link.
`
`Performing the primary backup to permanently-mounted random-access disk, rather than
`removable linear-access tape, allows an endless series of incremental backups while retaining
`a readily-accessible, full backup copy. As LAN disk space continues to grow, many sites will
`reach a point at which periodic full backups to tape are no longer practical. At this point, the
`ability to rebuild and restore entire lost hard disks in a reasonable period of time will be
`questionable because of the need to locate, mount and read large numbers of tapes.
`
`In order to
`A full-disk rebuild can only be accomplished quickly from a full backup copy.
`maintain such a copy, other LAN-to-mainframe solutions keep much or all of the LAN
`backup data on mainframe DASD, currently at 4 to 5 times the cost of LAN server disk
`storage. ESM keeps the primary backup data on PC hard disk, a much less expensive
`medium. Also, data available on local disk can be restored at LAN speeds (for example, 16
`mbps) while LAN data on mainframe DASD must be restored at WAN speeds (for example,
`56 kbps).
`’
`
`Advantages
`
`Full and incremental
`
`backups
`
`Primary on hard disk
`
`Minimal space,
`maximum benefits
`
`On a typical LAN, about 25 percent of server disk spaCe must be allocated to hold the local
`compressed backup files. ESM employs duplicate file elimination, data compression, and
`selective backup to minimize the storage resources required for local compressed backup.
`
`If only limited LAN disk space is available, ESM uses the available space to hold the files
`most important for completing a rapid restore - system files, a representation of the various
`disk's directory structures, and the most recently accessed user files. Other data is stored at
`the mainframe location.
`
`Mainframe security
`
`The mainframe provides a centralized data vaulting capability in the most secure corporate
`environment. Disaster recovery processes that provide for off-site storage of critical
`corporate data can be adapted easily to include data that is vaulted by ESM. Because data
`saved by ESM is automatically decomposed before it is transmitted to the mainframe , it
`cannot be reconstructed without the reconstruction recipe, which ESM maintains on the
`server, providing added security.
`
`3’2
`
`ESM Product Introduction
`
`
`
`
`
`Duplicate File Elimination
`
`ESM performs duplicate elimination at both the site and enterprise levels to conserve disk and
`tape storage and reduce the volume of data that must be transmitted over telecommunications
`lines to the mainframe data center. Duplicate elimination is inherent in ESM‘s exclusive
`BOBID-based design and requires no work by administrators. ESM's enterprise duplicate
`elimination can speed first-night backup of servers and LANs as more are brought under
`management by ESM.
`In addition,
`it significantly reduces processing and transmission
`overhead when new copies of applications software or data are distributed to a large number
`of users.
`
`for example, a new version of Word for WindowsTM is distributed to 800 servers
`If,
`throughout an enterprise, ESM ensures that only a single copy of the application is
`transmitted back across the network to the mainframe during the next incremental backup.
`
`ESM's duplicate elimination feature reduces network load in other ways. For example, in the
`process of reorganizing the directory structure, a user may create numerous apparently new
`files with the COPY command. However, regardless of how a directory structure is
`reorganized, ESM will recognize the presence of duplicate data under the new structure and
`will not back it up.
`
`A side benefit of ESM's approach to data backup results in reduced backup for some modified
`files. When file changes occur at the end of a large file, only the last portion of the file is
`backed up by ESM.
`
`ESM recognizes that the file is not the appropriate concept on which to base enterprise
`storage management, because the meaning of "file" varies according to the computing system.
`A UNIX file is not the same thing as an OS/2 file or a Macintosh file. Files can have many
`components - such as file block, extended attributes (EAS), or access control lists (ACLs) -
`which may change independently of the file data.
`
`Even within one computer system, such as 08/2, files can differ tremendously in size, making
`them difficult to process conveniently. A large enterprise can have millions - and, soon,
`billions - of individual files. The fully qualified names of all of these distributed files -
`site/domain/server/drive/directory/sub-directory/extension/filename — can be difficult or
`impossible to store in a central database. Also, a file name might not uniquely identify the
`data it represents.
`
`Binary Objects (BOBs)
`
`Chapter 3. ESM Features 3-3
`
`
`
`v c: \DATA\ESM. DOC
`
`'Instance
`Records‘
`
`C: \DATA\ESM . DOC
`BOBID! 1
`BOBIDIZ
`aoarma
`533332
`
`
`
`4-7.”?
`“mm
`
`The BOBID
`
`Figure 3-1. BOB Processing.
`
`ESM decomposes files into binary objects, or BOBs, that it subsequently handles separately.
`For a given file, the EAs, ACLs and data are stored and handled as separate BOBs. For large
`files, the data portion is broken up into a number of BOBs of 1 megabyte. These BOBs then
`become the unit of processing, transmission, and restart/recovery. Only the file data base on
`the server on which it resides has the recipe for reconstituting a file from its constituent
`BOBs. Of course, this file data base is itself backed up and is restored to a lost server prior to
`restoring the remainder of the backup data.
`
`Within ESM, BOBs are not assigned names based upon the name of the original file, an
`arbitrary numbering scheme, or other random criteria.
`Instead, the name of a BOB, the
`BOBID, is calculated from the contents of the BOB itself.
`
`All BOBIDs have exactly the same length, 128 bits (16 bytes). Each is equal to a 32-bit
`cyclical redundancy check (CRC) of the BOB, followed by a 32-bit LRC (longitudinal
`redundancy check, or XOR), followed by a 32—bit hash value, followed by 32 bits containing
`the length of the BOB in bytes.
`In cases where the BOB is 12 bytes long or less, the entire
`BOB (padded with zeros if necessary) is used as the first 12 bytes of its own BOBID.
`
`Although the BOBID is very compact, it is extremely robust mathematically. For example, if
`an enterprise had a single ESM site (generally no bigger than a single LAN) with a trillion
`BOBs (which would occupy millions of gigabytes of disk space), there would be less than one
`chance in a quadrillion that a file might be restored incorrectly because one of its BOBs had
`been assigned the same BOBID as a different BOB.
`
`BOB benefits
`
`The BOBID construct makes enterprise-wide distributed storage management practical in
`se