Handbook of Local Area Networks, 1998 Edition:LAN Management Click Here! Search the site:   ITLibrary ITKnowledge EXPERT SEARCH Programming Languages Databases Security Web Services Network Services Middleware Components Operating Systems User Interfaces Groupware & Collaboration Content Management Productivity Applications Hardware Fun & Games EarthWeb sites Crossnodes Datamation Developer.com DICE EarthWeb.com EarthWeb Direct ERP Hub Gamelan GoCertify.com HTMLGoodies Intranet Journal IT Knowledge IT Library JavaGoodies JARS JavaScripts.com open source IT RoadCoders Y2K Info Previous Table of Contents Next 7-4Network Data and Storage Management Techniques LAWRENCE D. ROGERS One of the fundamental services provided by a multi-user computer system has always been access to shared data. The shared data, once counted in megabytes, is now usually counted in gigabytes. Protecting and managing its storage is one of the most important tasks of the LAN administrator. To protect the data and manage storage requires a combination of hardware protection, operating systems services, and software tools. There are two main problems—maintaining data integrity and storage management. Data integrity means that systems data is always correct or can be corrected. Data integrity systems provide the ability to recover data in the face of hardware, software, and human error or from direct physical or programmatic attack. Error correcting codes (ECC) and fault tolerant disk systems can ensure immunity to all but the most disastrous hardware failures, and corrections are performed, for the most part, invisibly. Recovery from software overwrite failures and human errors, however, relies on keeping multiple copies of files and file histories. Disaster recovery (including recovery from direct physical attack) is usually achieved by keeping copies and file histories at separate secure locations. In the latter cases some human intervention is required, if only to make choices during the recovery process. Protection from programmatic attack requires an adequate security system and the use of virus detection or prevention programs with a file copy and history system for when serious security breaches are committed. Storage management means optimizing the use of systems storage media with respect to both cost and performance. Moving infrequently accessed data from expensive primary storage to less expensive storage media improves access performance and makes primary storage space available for the most commonly used data. Defragmenting disks improves performance by reducing access times. Compressing files frees disk space. Balancing the load across multiple disk drives and servers to prevent simultaneous over-utilization and under-utilization of disks in the same system smoothes performance. These and other techniques are commonly applied to storage systems to optimize performance and cost. This chapter discusses tools and techniques used for maintaining file histories and for managing the location of files in the storage system and briefly touches on related disaster recovery techniques. The chapter does not discuss fault tolerant systems or hardware technology for ensuring first-level data protection, nor does it discuss antivirus software or security or many other related topics. Each is beyond the scope of this chapter. MAINTAINING FILE HISTORIES Backup and Restore An important technique for ensuring that data is not lost is to maintain a version history for each file. This is done by periodically copying files to a different media, such as tape or optical disk. This is usually referred to as backing up files, because it involves making backup copies of files in case an original is lost or destroyed. The versions of a file available are determined by the frequency with which these copies are made and how long the copy is kept. The two main components of a backup strategy are the type of backup performed and the media rotation strategy. This chapter first defines some different types of backup and then how these can be combined with various rotation schemes. Full Backup A full backup (i.e., a backup of all files) made on a regular basis allows a network administrator to recover versions of all files from the system. The available versions of any given file are determined by the interval between backups. Periodic full backups have the advantage of containing all the systems data on one backup media, or at least a minimal number of media. This can help speed recovery if a full volume must be restored. Due to the size of a full backup, and hence the time required to perform one, it may be difficult or impractical to execute full backups as often as desired. Properly combining full backups with partial backups can still give reasonable protection. Incremental Backup An incremental backup captures all files that have been modified since the last full or incremental backup. It requires less time to perform because less data is being copied. To restore a full volume requires two steps, first restoring from the last full backup tape and then applying each incremental backup. Generally a full backup is done each weekend and incrementals each weekday. Differential Backup A differential backup captures all files that have been modified since the last full backup. This means that a Monday differential captures Monday’s changes after a weekend full backup. A Tuesday differential captures’ Monday’s and Tuesday’s changes. Differentials done later in the week require more space and time than incrementals, but a restore requires only the full backup restore and the latest differential. Modified Full Backup A modified full backup captures all files, except those for which three unmodified copies already exist on three different media. In other words, if a file has been copied three times each to a different media, and it has not changed, it will not be backed up again. This reduces the time to do a full backup because files that do not change are not backed up more than three times. If tape rotation schemes are used, precaution must be taken to see that needed copies are not overwritten. Modified full backups done on the weekend, with incrementals or differentials during the week, will ensure that the file history has copies of all files in their most recent state. Weekend backup time will be shorter, but the files will be spread across many different media. The big disadvantage is restoring a full volume or even a directory because there is a high probability the required files will be stored on many different tapes. Previous Table of Contents Next Use of this site is subject certain Terms & Conditions. Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.