|  | | Data Protection and Recovery | |
Data Protection and Recovery capability is a critical part of an optimised infrastructure. It provides a structure for disciplined backup, storage, and restoration of an organisation’s most value asset.
Although perceived as a largely “preventative” measure, effective Data Protection and Recovery is crucial to realising business goals. It does so by minimising downtime due to data loss, and preserving the productivity of both your IT staff and your end-users.
Advancing and organisation’s data protection and recovery capabilities proceeds through three stages: first, adding backup and restore functionality to critical servers; next, adding backup and restore functionality to all servers and implementing service level agreements (SLAs); and finally, extending the backup and restore functionality to protect desktop data as well.
| |
| | |
The three key challenges associated with data protection are:
- The recovery needs of the business versus the recovery capabilities in the protection solution.
- The distance desired between the production data and the protection solution.
- The protect-ability and accessibility of the various data sources within the environment.
Within larger organisation, IT is often isolated from the business units; this results in IT being unaware of recovery goals, and which data needs to be retained and protected. In medium-sized organisations, with fewer IT staff, this knowledge gap can be even worse.
Recovery Goals
In the past, traditional data protection usually took one of two extremes: either synchronous mirrored hardware or nightly tape backup.
With synchronous solutions, the disk was eliminated as a point of failure, since a second was always in sync, but at a cost that was prohibitive for all but the largest organisations. But even for those that have synchronous mirrors of their data, the data is not protected. Since there is only one logical instance, any incorrect writing to the file damages both physical iterations. In addition, a single server with synchronous disks still has significant points of failure that make both disk copies inaccessible.
The only other traditional alternative was nightly tape backup, which presents its own set of issues:
- Data was protected only once per day (in the middle of the night). Any data created during the business day, but then later overwritten or deleted, could never be recovered.
- While many companies perform backup in anticipation of a total server or disk rebuild, 97% of all restores are for single files or directories. This forces tape drives to rapidly spin forward and back repeatedly, in a process called shoe-shining, to locate the exact 1/8-inch of tape media that holds the single file to be restored. In general, tape is not an efficient restoration medium; tape is a safe long-term retention medium. And it takes a lot more time, from both an IT staff and end-user points of view.
- Tape restores of data has many associated problems, including corrupt indexes, mislabeled cartridges, unreadable media, recovery of off-site tapes, etc. The speed of tape restores is slow, which is one of many reasons why VTL (virtual tape libraries) allow faster disk to be presented as tape, for use by tape backup products. An improvement on this is D2D (disk-to-disk) which allows disk to be used directly, instead of misrepresenting as tape. This alleviates many of the other deficiencies of tape described earlier.
Distance from Production to Protection
Simply having redundant disk or tape backups does not satisfy most business recovery goals, mostly due to distance. For data to be adequately be protected, the redundant data, in whatever form, must be in a different geography and not susceptible to whatever potential calamity might affect the production server(s). If a building were to suffer any kind of flooding, wind, fire or other catastrophe, both disks and/or the tapes will suffer the same fate.
While a certain percentage of companies have offsite storage for their key datacenters, regional locations and branch offices often go unprotected. Up to 60% of corporate data exists outside of the data center, so even those “well-prepared” environments may only have 40% of what they need to resume doing business.
The "Protect-ability" of various data sources
In larger IT environments, protection and recovery nirvana is one solution that can reliably back up and restore every kind of data source. But even well-managed IT environments always have one or two applications that are challenging to protect, and burdensome to recover – if this is possible at all. For many, the best solution is to strategically align a few key protection solutions that can address a variety of specific workloads, often with a single back end.
This might mean one protection solution for Windows, with another for mid-range systems; or one solution for SQL and Exchange, and another for Oracle and SAP. Often, one of those protection solutions would have a tape back-end to protect the other.
|
|
| |
|
|
| |  | |  |  |  | Microsoft Server Client Access Licenses promotion for midsize businesses - a cost-effective way to purchase Windows Server 2003 R2 and Exchange Server 2003 CALs. | | Get a great deal now | | Register for a Microsoft Infrastructure Partner event to gain first hand knowledge. | | | | Order complimentary Windows Server 2003 R2 trial software. | | Order your trial software now |
|
|  | |  |  | | Register here to receive updates when new materials are available in your interest area. | | Register now |
|
|
|
