• 新聞中心
• 新聞稿
• 媒體資源
• 媒體聯絡
• 亞洲媒體話題
• 國際媒體話題
• 獲獎
• 網站地圖

電子報

CUES

May 23, 2007

by Tom Johansmeyer

Original article at CUES web site

Saving on Storage

Six considerations can make back-up and recovery systems more cost-effective

Despite what seems to be the perpetual uphill battle of IT expenses vs. executing on corporate finance priorities, it is possible to keep back-up and recovery costs under control. Even with rapidly growing storage requirements, you can manage the entire back-up and recovery process without committing more servers and people to existing processes.

Through the effective use of storage virtualization, an IT manager of any business operation, be it a regional credit union or a multi-national bank, can make back-up files portable across physical and virtual environments, reducing the time and cost involved in protecting data. A company can realize cost savings both through lower equipment costs and fewer staff hours for storage tasks, extending the benefits across all aspects of data center operations.

In the past, data and applications were tied closely to the hardware platform, requiring extra equipment that served no purpose aside from waiting for infrequent use. Storage virtualization shrinks the necessary server footprint for each system in the data center, requiring less staff backup and less planning coordination as well.

What Is Storage Virtualization?

Traditionally, managing backup and recovery involved physical-to-physical transactions. You had to back up from and restore to similar physical environments — such as one Dell server of a particular model number to an identical Dell server. Equipment was central to all back-up and recovery operations, with compatibility being the fundamental barrier to rapid task completion.

The next step in the back-up evolution was virtual servers (multiple servers that appear as one server), which offered limited success primarily because of the availability of device drivers. Ensuring that a recovery target had the correct drivers for the video card, network card, SCSI host controller or other connected devices required too much manual effort and resource coordination to deliver real process advantages. As a result, backing up to virtual servers was possible, but often not practical.

The latest innovation in back-up and recovery management is full storage and systems virtualization, offering complete portability of back-up files. Disk snapshots can be moved with relative freedom among physical and virtual environments with the right virtualization migration tools. This portability affords more flexibility and provides a vital key in a crowded data center and makes almost any available destination — physical or virtual — a viable back-up or recovery target. Some tools even make it possible to migrate from virtual back to a physical environment when the need arises, as it surely will as some applications, such as databases, grow beyond the limits of virtual servers.

Remember, these applications will be sharing the processing power of the server's central processing unit, so if you have one application on a virtual server that is hogging the CPU's power, all of the virtual servers on that machine will suffer reduced performance. The only way to deal with an application that is becoming too much of a CPU hog is to move it back to its own physical computer, where it will have dedicated hardware. Once that program is moved off the virtual server, all of the virtual servers that reside on that physical server will see improved performance.

Another key to virtualization is system state. You can think of system state as what a computer is like at a specific point in time, complete with all applications running, Windows running, and the user doing whatever he or she is doing. If you can preserve the system state when you take a snapshot of the hard disk, you have an exact duplicate of that computer. Let's say you restore that image, the computer will be in exactly the same condition as when you created the image — the same programs will be running, the same screen will be displayed and everything on that computer goes back to the very point in time when it was in a known, good working condition. Without restoring the system state, you effectively have just a file-based backup that can restore files, but does very little to restore the system to a working condition.

Effective back-up and recovery virtualization tools create full disk images in real time, even while Windows files are open and in use, opening back-up windows and making full backups possible at any time. As back-up and recovery virtualization tools have matured, disk imaging software that preserves the system state makes faster and more thorough back-up files that facilitate more accurate data and system recovery.

Cost savings thus result from:

• faster backups,
• storage flexibility,
• rapid recovery,
• transportable images,
• smaller equipment footprints and
• ease of use.

Faster Backups: Virtualized back-up and restore capabilities simplify both process-driven and ad hoc data backups through easier equipment provisioning, less need for administrator coordination and faster execution. Back-up file compression makes backups even faster with a smaller file footprint and less network strain.

Rapid Recovery: Data recovery is where speed matters most, as it implies that a system is unavailable. Storage media flexibility streamlines the recovery process, and full disk image snapshots allow you to recover to nearly any available server. The compressed back-up file accelerates recovery through network transmission speed and the time needed to write the file to the recovery target.

Storage Flexibility: More efficient backup and recovery operations result because data can be stored to or recovered from a broader variety of physical and virtual servers. Administrators spend less time securing a specific server or environment as more targets are available. Make sure to use tools with a library of device drivers built into the software, as these tools will mitigate the need for manual intervention, and full disk imaging during the backup process makes "bare-metal recovery" possible.

Transportable Images: We talked about storage flexibility; transportable images give you target flexibility. A transportable image, such as those from disaster recovery companies like Acronis Inc., can be restored to any hardware, regardless of the hardware's motherboard, CPU manufacturer or system design; the only limitation is the system must be x86-based, and running a version of Microsoft Windows. By eliminating the need for storing and maintaining identical hardware that is seldom, if ever used, a company can instead spend its resources where they're needed most, be it for additional hardware, staffing or non-IT expenditures.

Smaller Equipment Footprints: Faster and smaller back-up files do not tie up your networks or require extra servers, and flexibility with storage media makes it unnecessary to carry extra equipment to serve as a backup or recovery target. Easier provisioning leads to a smaller platform footprint. Lower equipment expenditures yield ongoing savings through lower maintenance costs on servers and software licenses.

Ease of Use: Fewer errors and less administrator intervention help you do it right the first time. Backups become more accurate and recoveries faster. Further, costly initial training and ongoing refresher sessions are unnecessary. Emerging back-up and recovery virtualization tools use intuitive Web-based interfaces to help administrators learn the software quickly, making the software useful more quickly and accelerating the realization of cost savings.

These six dimensions result in a lower total cost of ownership for all back-up and recovery operations, driving cost savings for every application and platform. Back-up and recovery virtualization is both inexpensive up front and more effective on an ongoing basis. Regardless of how a company goes about implementing its virtualization strategy, remember that recovery and being able to deploy an image to a new system is everything; having a back up — be it file-based or image-based — that cannot be deployed to a new system is worthless.

Tom Johansmeyer is a free-lance writer based in New York. Tom has written extensively on a variety of technology subjects and has implemented systems as a project manager with an international consulting firm.