As the real-time operational nature of active data warehousing evolves and the volume of data increases, data availability becomes critical to organizations making real-time business intelligence (BI) decisions. The increasing movement for these businesses to capture detailed data from more areas of the operation and for saving data for extended periods exemplifies the need to safely store and maintain that data.

enlarge Disk backup products offer the reliability of disk systems to provide predictable backup and restore time windows.

Regulatory compliance requirements mandating the long-term retention of data also make archiving increasingly imperative. Sarbanes-Oxley, the Health Insurance Portability and Accountability Act and the European Union Data Protection Directive, among other compliance regulations, are extending the need for long-term data retention—sometimes for as much as 10 years.

This all means that precautionary safeguards against unexpected data loss must be in place along with the means for storing and maintaining data to adhere to regulations. To ensure an organization's data is safe and readily accessible, it must be routinely backed up without causing unnecessary or excessive downtime. This adds up to a greater need for high-performing backup, archive and restore (BAR) solutions, especially for organizations with business-critical and large data repositories. BAR solutions must offer greater database availability, recoverability and protection in today's data warehousing environments.

Backing up to go forward
Backing up data is no longer just a minor housekeeping chore but a critical strategy. With the amount of data growing exponentially and with the increasing use of the data warehouse for real-time operations, business continuity is critical, and the recovery of data during loss incidents and data security is more important than ever. Without current or accessible data, companies cannot continue to function well.

Besides, data unavailability is costly—the longer a company is without its data, the more detrimental the effect. If data is lost, re-creating it and rebuilding data repositories from multiple sources is resource- and time-consuming and reconstructing BI data is difficult. An effective BAR solution, therefore, must be assessed as a business-critical function.

With an effective BAR solution, business continuity is ensured with these capabilities:

	Data protection. Preservation of data off-line
	Recoverability. Rapid restoration of data to its full integrity in the event of user errors, disasters, hardware failures, etc.

Teradata BAR partners and provisions > BakBone's NetVault is highly integrated with the Teradata platform. > Symantec's NetBackup is an integrated solution that is a standard industry platform. > Sun Microsystems is the industry's leading tape backup vendor. > LSI provides disk storage products. > EMC provides the virtual tape library appliance that uses disk but emulates most standard tape libraries and drives. > Protegrity provides a robust encryption solution for backup, archive and restore processing.   —R.L. enlarge Teradata partners with six technology suppliers to produce an industry-leading BAR solution.

BAR must also conform to system recovery objectives, which are predetermined measurements of the amount of time and age of the data the company can afford to be without during a system failure. They are defined as:

	Recovery time objective (RTO). Time and service level limits to which a data warehouse and its data must be restored to avoid the unacceptable consequences associated with extended breaks in business continuity
	Recovery point objective (RPO). Data's age, or the point in time to which that data must be restored, before its loss or the need to re-enter it is considered intolerable to users

To clarify, a company might establish an RTO of 10 hours (the amount of acceptable system downtime) and an RPO of 24 hours (the oldest acceptable age of the data).With an RPO of 24 hours, the system must be backed up every 24 hours so if it loses data and has to be restored from the backup, users will have access to data that was created no longer than one day prior. With an RTO of 10 hours, the data restoration must be accomplished within the 10-hour time window of the objective. Both measurements help determine the optimum configuration of the backup and recovery solution for a company. These objectives then guide the selection of the actual BAR solution components:

	Hardware. Tape and disk storage devices
	Software. Storage management and encryption
	Services. Planning, implementation and support of a system

Storage devices
Organizations have several choices in backup storage devices: Data can be backed up to tape, conventional disks, virtual tape libraries (VTLs) or, preferably, using a combination of tape and disk technologies. The decision is based on a number of factors, including:

	The amount of data and frequency of backup
	The time window in which data must be backed up or restored
	Accessibility of archived data
	Off-site or on-site storage capacity

For instance, a company with long business hours and an actively used data warehouse will have a smaller window in which to back up the data. If this company has a 24-hour RPO, the appropriate BAR solution must be one that enables a daily backup with smallest impact on database access.

The following describes the different types of storage devices available to users:

TAPE, THE MAINSTAY
Tape backup devices are the traditional and most widely deployed solution for data protection. As the device with the lowest cost per data block, tapes are typically used for long-term archiving of data. Their mobility makes them easy to store in a remote location, which offers a safe, cost-effective data refuge in case of a disaster.

Still, relying solely on tapes is an increasingly risky proposition. The dependability of tape-based backup and restore time windows poses a real challenge and is the driving force behind the adoption of disk backup.

DISK EMERGES
While tapes used to be the only option for backing up data, disk products evolved and became an important part of the solution. Their predictable backup and recovery times make them more adaptable to businesses with longer operational hours, thus shorter time windows for data backup. Disk systems are also considered more reliable than tapes because they do not wear out from steady use, nor do they suffer from media errors or mechanical problems.

This immunity of disk systems against media failures is gained through the use of redundant-array-of-independent-disks technology. This technique backs up data in several stripes either on the same disk or on multiple disks. With this redundant method, backup data is protected from any data mishaps on the disks that would have caused a backup or restore delay.

Two industry approaches used to back up data to disk-based storage are:

	Direct to disk. Standard disk files that directly store data
	VTL. Disk-based backup and restore appliance that emulates open systems tape libraries and easily fits into existing tape backup environment without changing current operations or infrastructure

The preferred BAR system configuration creates a copy of the data on a backup-to-disk device, then archives that data to tape with encryption for security (see figure 1 above). Typically, two copies of the data are kept on the disk backup devices, and two others are on tape for archival purposes. This approach achieves the best level of protection for users and can reliably meet aggressive performance goals such as a five- to six-hour backup and a 10- to 12-hour restore.

BAR for Teradata data warehousing
Teradata's massively parallel processing (MPP) architecture puts unique demands on a BAR solution. With MPP, multiple workloads are run in parallel on multiple nodes. This high-performance characteristic and high-speed data flow enable Teradata to quickly query large data tables. Because Teradata tables are hashed across all nodes and storage devices in the system, the parallel architecture puts each element of the BAR solution in the critical path to full performance.

Teradata offers a BAR solution strategy that makes BAR tasks easier, faster and safer. Teradata's BAR solutions are an effective way to leverage the performance and reliability of the Teradata data warehouse for protection of key data to ensure its safety and availability, and to meet backup and restore window needs. The scalable, packaged solution is optimized for high performance by leveraging the MPP approach and therefore minimizes backup and restore windows. Additional security is enabled with the safeguarding of data encryption added to disk and tape backup capabilities.

Putting BAR together
Management of the BAR solution leverages industry-standard backup application software that provides ease of use, operational performance and compatibility with data center operations.

Each of the three BAR management software alternatives (see figure 2, above) provides scheduling and management of backup and restore operations, control and monitoring of disk and tape, and movement of data between the Teradata system and BAR solution. The selected backup application for management links through a software access module called the Teradata Extension (see figure 3) to the Teradata Database nodes.

enlarge The nodes in the Teradata data warehouse are networked to the BAR framework servers running the BAR software (extension and application) to support disk and tape backup devices.

Teradata has forged partnerships with six technology suppliers to provide a fully integrated set of industry-leading, matched and certified components. These Teradata BAR elements represent best practices for backup and recovery configurations and deliver the best return on investment (ROI) for companies. Teradata integrates and certifies these components into a complete, end-to-end solution that includes one-stop implementation and support services.

Because tapes often contain sensitive data and are inherently portable (and therefore prone to misplacement or loss when moved off-site), backups that are stored on tapes are especially vulnerable to security breaches. Teradata Defiance BAR Encryption software protects this tape-based data by effectively providing a strong encryption capability. Through a selection of industry-standard encryption algorithms and the secure management of the location where encryption keys are stored and maintained, the backup tapes are well protected. Companies can be assured that their data, so crucial to the success of their business, is safely backed up and stored.

Taking control through BAR
Backing up and restoring data are vital functions for organizations that understand the importance of real-time data to help drive business success. Active data warehouses are a key element in business operations as leaders use this data for near real-time decision making.

For this reason, and to comply with numerous regulatory requirements, backed-up data must be kept available, safe and readily accessible. With Teradata BAR solutions, data can be quickly backed up and safely restored, thus enabling the business to continue to rely on its data for innovative decision making. T

Success is not fleeting A large financial institution in the United Kingdom is a long-time Teradata user. The organization started with a modest-sized Teradata system that provided decision-support workloads for back-office business workers without any real availability needs. These workers met their initial backup needs with a mainframe solution in which all backup and restore operations were performed through channel connections to mainframe tape devices. Over the years, the company's Teradata system became more mission-critical and hosted an increasing amount of business data. Yet the backup, archive and restore (BAR) strategy remained the same and the risk around data protection silently grew. A number of challenges demanding change were identified, such as enhanced recovery-time-objective and recovery-point-objective parameters and channel capacity shortfalls. Situation The mainframe approach required 39 hours to back up all important data. Yet the company had no more than 12 hours available, a time span that was predicted to shrink during the next 12 months. Resolution The financial institution implemented the Teradata BAR solution and achieved its goals, which included: > Dramatically improving backup and restore performance > Reducing costs by eliminating use of the mainframe infrastructure > Putting management of the data protection directly in the hands of the data warehouse staff > Building a scalable BAR approach aligned with Teradata for growth The new BAR solution provided backup of all the company's data, including 14TB of user data space, in a three-hour window. Now the company can perform all backups in an acceptable window of time. Just as important, it has removed the risks associated with not protecting all of its important business and customer data. —R.L.

Richard Leatherman is a former Teradata employee who now works as a freelance and technical writer in San Diego.

Jim Dietz, Teradata platform marketing manager, has more than 12 years of experience with Teradata in developing and marketing data warehouse solutions.

Teradata Magazine-December 2007