【正文】 for a defined window of time allocated to backups. However, many organizations are finding that they can no longer afford to have routine designated downtime during which backups must be run. In addition, the amount of time allocated for server and data recovery is shrinking because every minute of downtime is being more expensive. ? Branch and remote offices need equal protection. Centralization and consolidation have recently bee buzzwords in the IT industry, but the fact remains that a great many businesses have branch or remote offices whose operations and resources cannot be easily or feasibly consolidated. Examples include retailers, financial services panies, and manufacturers. The resources in these branch offices are often as important as the organization’s centralized resources, but branch and remote offices typically don’t receive the same level or quality of data protection because implementing it is expensive and plicated.? Cost control is driving vendor consolidation. One way that businesses have identified to lower IT costs is to reduce the number of vendors with which they do business. This trend increases the pressure on panies to reduce the plexity and operational cost of their data protection infrastructures by reducing the number of solution vendors in their environments.The emergence of these trends means that it is now possible to make a strong business case for deploying data protection systems that help meet these business challenges. Of course, there are technical challenges inherent in designing a data protection system that will effectively protect enterprise data while meeting these business needs.Technical Obstacles to Effective Data ProtectionMany of the technical constraints that hamper data protection effectiveness today actually derive from the nature of tapebased backup systems, media and processes. Understanding what these constraints are is a necessary part of designing a system that provides better protection while being responsive to the business requirements described in the preceding section.Tape System Efficiency, Robustness, and CostTape systems have traditionally offered what seemed like a reasonable tradeoff: low media cost (and thus low longterm archival and storage costs) versus limited performance. As a mature technology, tape systems are familiar to most IT professionals and decision makers, and virtually every major operating system and backup solution provides support for tape devices. However, using tape as the primary backup and recovery mechanism also imposes constraints.? Tape systems are very slow pared to diskbased solutions, both for backup and recovery operations. This speed gap has widened as disk storage systems have increased in speed and I/O bandwidth. Requirements to back up more data faster often e into direct conflict with the speed limits imposed by tape systems.? Tape systems include a relatively large number of moving parts。 the electromechanical ponents that physically move the tape are prone to failure at higher rates than solidstate ponents. The tapes themselves are subject to physical wear and must be stored and maintained in the right environmental conditions to remain usable. This reality is often underappreciated when one is relying on longterm viability of data.? Tapebased systems don’t provide an effective means of centralizing backup and restore processes. Most tapebased backup solutions require either a hefty allocation of bandwidth between a remote site and the central backup site or a local tape drive at the remote site, which then introduces the problem of how media are managed, cataloged, and stored between the remote and central sites. ? Tape systems have historically been poorly integrated with diskbased backup systems. Companies seeking to bine the high performance and robustness of diskbased backup with the low longterm storage costs of tape solutions have frequently found that bining systems from multiple vendors gives them all the drawbacks of both methods. In consideration of these constraints, more and more panies are recognizing that tape is not a preferable medium for routine data recovery, but is best served for longterm data retention and archival where disk may not be as practical.Network Bandwidth, Latency, and UsageBackup systems operate by making a faithful copy of a set of protected data items. To do this, they must be able to read and copy all the data from the source data items, then transmit the copied data to the location where it will be written to the backup medium. In environments where all backups are done locally, this can be reasonably straightforward. However, the more mon case is also more plicated: when it bees necessary to back up data from one server and ship it to another server to actually perform the backup. Conventional backup systems operate by making wholesale copies of every bit of the source data. This is initially required for any backup system. However, performing plete copies on a routine basis uses a large amount of network bandwidth to move the copies from the source server to the backup target. This problem is exacerbated in environments with limited available bandwidth, high latencies, or poor network stability. Application Awareness and SupportNot every application is created equal. Some applications create and process “flat” files that are opened, changed, and then closed (. Microsoft Office documents)。 in addition, you can schedule tapebased backup jobs to capture regular full backups to tape to meet your archiving and pliance needs while still preserving your ability to do finegrained restores at high speed directly from disk. In addition to directly restoring individual protected databases, you can also use DPM to capture system state data so that you can restore an entire protected server. These restores can use any of the past iterations of data you’ve chosen to capture on the DPM se