|
By George Symons, CTO of Information Management at EMC Corporation.
Intelligent data protection in the enterprise means designing backup, recovery and archiving procedures around the service level requirements of the application and the policies of the organization to meet the performance and availability needs, while extracting business value and ensuring that compliance
requirements are met for all data.
Drivers and strategies
Compliance and business value are the two primary drivers changing how data is managed and protected in an enterprise today.
Compliance
Compliance is a broad term and covers many activities such as regulatory compliance, industry compliance, corporate governance, ediscovery and litigation support. The enterprise faces two major challenges around compliance. IT must be able to retrieve requested information and it must get it done within the time constraints. A proper compliance strategy includes information lifecycle management (ILM) functions such as archiving data onto disk for efficient retrieval, and using policies to move compliant data onto long term storage with retention capabilities.
Business value
More and more data that is stored on disk is actually business information. The value of this data can be immense. Value is not only in structured databases, but in semi- and unstructured data, such as e-mail, word processing documents and spreadsheets.
Intelligently managing valuable data across the enterprise requires data classification to understand the value of the data in individual files, so that it can be protected, archived and retained in a means that ties the storage management policies to the requirements of that data within the corporation.
Intelligent data protection: short-term and long-term
Tiered storage is the solution for managing stored data on the optimal cost / performance storage device based on the data’s value, retrieval and retention requirements throughout its lifecycle. Individual pieces of data are treated differently according to its contents, associated application, stage of life, compliance requirements and business value. The IT organization also needs to look at the different requirements for operational recovery, disaster recovery and archive. The majority of recovery requests are for files that have been deleted or corrupted (operational recovery) and has been protected within the last 24 to 48 hours, this data should be kept on disk for fast recovery. Long-term storage strategies are for disaster recovery and archived data that must remain protected and accessible for long periods of time, but will not require immediate user access.
Keep 72 hours worth of data on disk
Operational recovery, which is the individual file or object recovery done because of deleted or corrupted objects requires newly backed up data to be kept on primary backup disk for at least 72 hours. This will dramatically reduce recovery times. Once data ages it can be moved to online or offline lower cost storage depending on its nature and access needs. The more critical the data, the better suited it is to online storage.
Reduce the recovery point objective
Most companies backup in 24-hour intervals, but this schedule can conceivably lead to 24 hours worth of data loss. This is where snapshot technologies come in with their ability to restore to more recent points in time. Depending on the nature of the data, snapshots can bring the recovery point objective down to nearly zero through new technologies providing continuous protection. Not all applications need this level of recovery and for some, a 24-hour recovery point is perfectly acceptable. But the more critical the application and the more often data is modified, the smaller the recovery point objective should be.
Use disk as the primary initial target
Some companies have adopted virtual tape libraries, which have the advantage of backing up at disk-based speeds without having to change tape backup procedures, policies and applications. However, virtual tape libraries store tape images, which negatively impact individual file recovery. Changing over to true disk backup will take an initial investment of money and time, but disk’s tremendous advantages trump any momentary inconvenience. By backing up to disk (not virtual tape), companies are able to use disk’s considerable recovery advantages, including simultaneous multiple operations like reads and writes, faster access and granular recovery options instead of scanning through tape images.
Long-term: retention over the long haul
Practice disaster recovery
Corporate IT should architect disaster recovery procedures according to data’s value and compliance requirements. Critical data might be replicated to a disaster recovery site or backup media might be physically moved off-site at frequent intervals. Whatever the choice, DR operations must be matched to the data’s recovery time. Data could be restored in more than 24, 48 or even 72 hours without consequences, but some data must be immediately restored. This is why backup media management is critical: IT must know exactly where the most efficient copy of the data is. The data could be on local disk or tape in a library, or replicated to a disaster site, or on tapes in an offsite archive,
Add archive to DR
Even large businesses have historically used backups as archives. However, backup and recovery applications are not a reasonable method for long term retention and easy access of data in later discovery operations. Backup/recovery is about recovering individual files, directories or volumes and tends to keep many copies of the same files. Archives on the other hand are to be searched and need to be accessible (not recoverable). Also, maintaining the retention in an archive allows files to be deleted when their expiration time arrives rather than managing the tape life of an entire file system containing data of differing value and retention requirements.
A popular investment bank recently experienced the crippling cost of using backup tapes as archives. The investment bank repeatedly failed to produce e-mails in response to a broad discovery motion. Producing the e-mails required searching through thousands of tapes. This was bad enough, but then after IT executives had insisted they had searched all their tapes IT staff found three separate caches of backup tapes in three different locations. The judge was not amused and fined them more than a million dollars for the failed search.
Best practices for the enterprise
Classify data to establish data management policies based on service level requirements. Classification includes disaster recovery requirements, point-in-time recovery objectives, retention periods, and so on. Data classification and assessment across the enterprise is a large project, and professional services organizations with assessment and tiered storage expertise can be invaluable.
Use disk technology for applications that have significant service level agreements (SLAs) around recovery point and time objectives.
Separate the concepts of backup and recovery from archive. Use backup and recovery for disaster recovery and operational recovery, and archiving for lifecycle management, retention periods, discovery and compliance.
Clearly understand what your compliance requirements are government, industry, corporate governance or litigation. Most enterprises must be prepared for all of these types, and corporate IT should build storage management strategies around differing compliance requirements and discovery motions.
Planning and deploying proactive data protection in the enterprise can be complex. Consider bringing in consultants who have built successful practices around backup and recovery and archiving in enterprise environments. The consultancy must be able to carry out an enterprise-wide assessment across the entire infrastructure: understanding both storage and data networks, storage technologies and application requirements.
It is important to understand the requirements of the lines of business, the IT organization, the general counsel, and corporate security. Successful assessment, planning and deployment will allow the enterprise to architect a proactive data protection solution. This solution will enable the company
to not only survive data challenges like loss and broad discovery motions, but also to strategically manage data for increased profitability.
http://www.emc.com/
EMC is exhibiting at Storage Expo 2006. Now in its 6th year, the show features a comprehensive free education programme and over 90 exhibitors at the National Hall, Olympia, London from 18 - 19 October 2006 www.storage-expo.com
•Date: 14th September 2006 • Region: World• Type: Article •Topic: IT continuity
Rate this article or make a comment - click here |
