Tape backup: the cornerstone of data protection.

The power blackouts that hit the Northeast U.S., Canada and the U.K. this past August was a wake-up call to organizations of all sizes that solid data protection processes must be in place, not only to survive these events, but also to maintain a competitive advantage.

Executives and IT directors working for companies affected by the recent blackouts need to ask themselves a couple of questions. First, could they have gotten back up any quicker? Second, were any of our competitors back in business before us?

Among the heroes of the recent blackouts are the IT staff at the New York Stock Exchange, which opened on time Friday, August 15, just a couple of hours after power was restored in lower Manhattan. The majority of the IT staff spent the entire night at the NYSE, unable to call home or connect with family and friends. Despite all the technology processes and disaster recovery plans, it was the people who made the difference. Thousands of professionals like those at the NYSE who ensure data is always protected are the people that contributed to the U.S., Canada and the U.K. getting through these events with remarkable little disruption. But of course, not all businesses were back in business as quickly as the NYSE.

Data protection and business continuance programs encompass any number of techniques, depending on the size and type of enterprise, the types of applications, the type of data being protected and the organization's expectations.

For example, the most effective measure to protect against power failure are Uninterruptible Power Supply (UPS) systems to provide time to shut down systems in an orderly fashion, and backup generators to provide electricity to essential systems such as medical equipment, air flow and water pumps. But a UPS is a very specific and short-term patch: only the most aggressive organizations can afford to have enough backup power to run their data centers through a day-long outage. UPS protection is just a part of a wide range data protection and business continuance program.

The ongoing threat of terrorism has led many companies in New York to set up mirrored data center sites (also called Remote Replication) in remote locations, such as New Jersey, to protect their data resources from potential terrorist attacks. But in the case of last month's power failure, these sites proved less adapted to maintain business continuity since the lights simultaneously went out in New York and New Jersey. Moreover, an unstable power supply can often corrupt data, and mirroring doesn't protect against corrupted data. If the primary data store becomes corrupted or erased, so does the mirror. And viruses present a similar danger of corruption of both primary and mirrored data. Remember, remote data replication is only one component of a deeper data protection strategy.

Backing up data to tape is the cornerstone of almost every data protection program. Tape provides fast data transfer performance, high capacity and reliable recovery at an extremely attractive cost. Implementation of established tape backup techniques allow a data center to be brought back to a known-good state, eliminating any corruption or viruses that may have affected the system. And in a true disaster recovery situation, the key to tape's use in a true disaster situation is the ability to remove it from the system and store it safely and inexpensively offsite and/or move it (and therefore your data set) to a new site if needed. Tape backup is also more immune to power outages than other technologies, such as disk systems, since backup operations are usually done late at night when network traffic and power demands are at their lowest.

Understanding Common Data Protection Strategies

As tape technology has evolved over its 50-year history, there are a number of sophisticated strategies IT staffs have found useful in building foundation for data protection and recovery.

Some organizations have incorporated a concept called tape pipelining that overcomes the latency problems that were associated with performing backups in a remote facility--even one that is hundreds or thousands of miles away from the primary system--by sending the data over standard IP-networks. Performing tape backups remotely eliminates the need to manually remove tapes from libraries and transport them to the remote facility, and at only a slightly higher cost than a local backup system.

In more sophisticated networks, where downtime is measured in millions of dollars, point-in-time-copy is a useful technique that delivers fast access to a data backup. When using the point-in-time strategy, backups occur more frequently than the traditional once-a-day backups. The data is typically staged to hard disk (prior to writing to tape) so recovery of lost data can be accomplished quickly. This requires a high-level of network bandwidth and a further duplication of storage resources (similar to mirroring). Therefore, point-in-time-copying is usually more appropriate for larger organizations that can make and sustain the significant strategic investment of maintaining frequent data backups throughout the day.

Another common method for limiting the effects of an IT disaster is data archiving--from simple copy-and-purge operations to Hierarchical Storage Management (HSM), Content Management, and Information Lifecycle Management (ILM). Increasingly, archival storage means moving large amounts of older, infrequently accessed data to tape. As a large percentage of organizations are using automated tape systems to offload their primary storage servers--a majority of their data assets--to tape media, IT managers can now expect professional-grade tapes to reliably retain data for at least 15 years. And when tape archive systems are connected to the network, all data remains accessible to the user base--albeit at slightly longer latency times--while everything else runs faster and smoother because so much of the network traffic associated with data storage is eliminated.

Most archival software applications make it a simple, automated task to make copies of archival media for off-site storage to protect this important data from fires, floods and other localized disasters. And of course, a tape sitting on a shelf in a vault somewhere is not affected by a power outage the way a spinning hard disk would be.

Congratulations should be awarded to the IT staff at the New York Stock Exchange, for opening on time on Friday, August 15th, just a couple of hours after the power came back on in lower Manhattan. Most of these people spent the whole night at The Exchange, unable to call home or find out about loved ones, catching just quick naps in conference rooms. It's very important to note that despite all the technology and all the plans, it is the thousands of professionals like those at the NYSE that make up the backbone of any disaster recovery.

Data Protection and Business Continuity: A New Competitive Advantage

So, back to the questions. Who was up and running faster--you or your competitor? The costs of the 2003 blackouts, if indeed we've seen all of them already, will be very hard to calculate. The costs to individual businesses start with the revenue opportunities that they missed while the lights were out and the ensuing time that they spent getting their systems back up. But the true costs will be known only after a couple of quarters, once their customers have a chance to evaluate how these businesses reacted, bow they performed, and how they protected their data through these events. The company that was able to recover quickly will keep its customers--and may pick up more from the company whose records were inaccessible during the critical days of recovery. Customers may possibly decide to take their business elsewhere--to someone who is better prepared to handle inevitable future IT disasters.

Two things are clear from this "wake-up call": every IT operation needs to have data protection and business continuance plans and the people and equipment to properly execute them; and the use of tape technologies continues to be an essential, cost-effective component of these plans.

www.tapecouncil.org

SOME DATA SETS ARE MORE EQUAL THAN OTHERS

Almost all data that an organization generates and stores can be considered important to its continued operations. However, some data and applications have greater priority than others, which should be taken into account when designing backup, disaster recovery and business continuance programs. When assessing the hundreds of software applications that an enterprise may support, TTCouncil suggests asking the following questions in designing an efficient data protection system:

* Is the application and/or its data essential for getting back in business quickly following an outage'?

* Is a local or remote mirror needed or appropriate for quick recovery following an equipment failure?

* How long must the data be retained, and does it become a liability when it is no longer needed (e.g., in legal proceedings)?

* Is the data subject to regulated retention and auditing requirements, and does it need to be stored on non-alterable (write once) media?

* What is the user access time requirement for the data, and can it be moved to a near-line automated archival system for more cost-effective long-term storage?

* Will you keep both local and remote backup tapes to cover both accidental data loss (local recovery) and IT disasters (remote recovery)?

* Does the remote copy need to be on-line for last recovery, or will an off-line vault provide sufficient recovery time (minutes-to-hours vs. hours-lo-days)?

* For long-term data storage, what provisions are needed for having compatible hardware and software to read the data years after it has been created? Or do you need to plan on periodic data conversions to newer storage technologies?

Richard Harada is president of the Tape Technology Council, the primary industry resource for promoting, integrating and using tape storage devices.