Computer back-up tapes are under attack. Scan recent news headlines and advertisements in the storage industry, and you'll see tape is blamed for everything wrong with backup and archiving. What's more, many IT managers and executives increasingly list "eliminating tape backup" as one of their top initiatives. And, there's a long list of technologies (and associated acronyms) aiming to replace tape: Virtual Tape Libraries (VTL), Wide Area File Services (WAFS), de-duplication, Disk-to-Disk (D2D) backup-just to name a few.
Despite the apparent movement against tape, organizations of all sizes continue to use tape effectively. These organizations understand how to leverage tape's substantial strengths and avoid its weaknesses. By combining online (disk) and offline (tape) media, organizations can enjoy both cost and performance advantages.
So why has tape lost its luster? For most, the answer lies not with the media itself, but with the process. While tape and tape-drive vendors have improved the performance and capacity of their products to meet IT storage needs, the software and processes behind tape backup have not adapted as quickly.
Most of the headaches associated with tape originate from a company's computer backup software and processes. Backup software operates assuming that recovery time and reliability are the primary objectives. This orientation primes the company for problems because optimizing recovery and maintaining an effective archive require separate processes and approaches.
What most don't realize is that the assumptions that created the traditional backup paradigm have changed. For example, as companies have increasingly kept information for compliance, business intelligence and other reasons, they've relied on backup tapes as their chief retention tool. The trouble is that backups reflect the state of systems at the time they were created and, as tapes sit in storage over time, the data on them tends to become more difficult and expensive to access. This is because the platforms, applications and people regularly change. Accommodating these changes can be difficult.
Let's take a step back and explore the situations where organizations might use tape today.
Recovery from System Failures
The goal when recovering from hardware failures is to lose as little time and data as possible. Meeting this goal for the most critical applications requires various types of replication and clustering, or other processes that offer high performance and are on-line or near line. Virtualization technologies have helped by reducing the likelihood of needing to recover and making recovery less dependent on specific hardware. Tape backups can serve as a fall-back if the online methods fail, and tape backups can also serve as the alternative for less critical systems. Some organizations are turning to VTL or other disk-based backup approaches as the first tier of backup for recovering from these types of failures. These online approaches to recovery generally offer faster times and storage efficiencies through reduction of data redundancy.
Recovery from Site or Regional Disasters
Unlike system failures, site or regional issues provide a much greater set of challenges. All of the organization's assumptions on recovery go out the window. You can't depend on having the same people or access to the same equipment. You also can't depend on the availability of outside partners like your local systems integrators. And while off-site replication may provide excellent recovery metrics (like Recovery Time and Recovery Point Objectives), it is vulnerable to the unknowns of any disaster. Often, when disasters occur, well-intentioned people make mistakes - like trying to manually fail-over systems while a disaster is happening and getting something wrong. Testing disaster scenarios can help prevent this; but ultimately, off-site tape backup is a highly effective failsafe when primary recovery systems go down. As a portable, off-line media, tape is excellent insurance against site or regional issues.
Recovery from Corruption-Related Failures
To recover from viruses, malicious users, common user errors, database corruptions and other threats, you need a solution that maintains a relatively short history of the data on your systems. Just a few years ago, almost all organizations used tape backup to recover from corruption. But as organizations have adopted newer technologies like snapshots, CDP and VTL, they've reduced the use of tape in normal restores. For the same reasons that tape is important for site-level disasters - its isolation from the primary system failures - tape backup allows companies more points in time to roll-back to and provides a secondary recovery mechanism if the online approaches fail. For these types of issues, a month or so of historic tape backups is generally sufficient and is very cost effective.
Long-Term Retention of Information
Years ago, IT organizations determined that backup tapes contained all the information in a given environment and that keeping the backup tapes would address the need to retain that information. However, tapes written by backup software for backup and recovery purposes are not ideal for retaining archival data. Here's why:
- Backup technologies must remain available to access the information. As media ages, the tougher and more expensive it becomes to access the data.
- The monthly or quarterly snapshots provided by backup tapes may not contain the right information. For example, an email received and then quickly deleted often won't make it onto a month- or quarter-end tape set.
- Data of varying importance and with ranging retention schedules is often combined on the same media to speed back-up and to conserve tapes. This practice results in keeping too much data for too long, which is expensive and a security liability.
- The backup process creates tremendous redundancy. Because only a small percentage of data changes between backups, the majority has already been backed-up.
These challenges are not problems with tape itself; they are flaws with the process and how the technology is used. With a few changes, companies can use tape more effectively for archiving and backup. Here are a few suggestions for your organization to better leverage tape's strengths:
- When discussing backup and archiving, change your mindset from tape versus disk to online versus off-line. Tape is a better off-line medium, and disk is a better online medium.
- Consider separating backup and archiving in process and technology. There are a number of Hierarchical Storage Management (HSM) and archiving solutions available for specific applications.
- Find ways to isolate fixed, or less active content, from active information. Many organizations have a large percentage of their data that needs to be retained, but that isn't accessed very often and usually doesn't change. If this data can be moved (virtually or physically) to different systems or to different backup schemas, the results can be stunning - less data to be backed up every day improves performance and reduces costs.
- If your backup technology supports it, create separate backup jobs for retention purposes and perform differential backups across those sets. In other words, don't just use the weekly full copy that happens to fall at the end of the month. Instead, create a separate backup job that runs differentials across monthly archive copies. Continue to run your normal backups with a 30-day rotation cycle. This will require that you run a separate job once a month, but will reduce the amount of archive media stored. Because these backups aren't about quick recovery, it doesn't matter if a restore takes more time across more tapes.
- Find a good third-party, data restoration vendor to help you organize and restore historic archive tapes. Although initial fees may seem high, the overall cost is often much less, the reliability is generally higher and the good vendors can offer advice on how to minimize long-term costs.
- Consider technology that allows you to aggregate and maintain your historic backup catalogues without having to keep the old backup servers and equipment running. When combined with 4 and 5 above, this will help you find the tapes you need to restore and provide a mechanism for restoring this information. This will likely increase reliability and reduce costs.
- Consider using a third-party to consolidate historic backup jobs on older media and backup formats. While you might reduce your storage costs, the bigger advantages are that you'll get a clearer picture of your inventory and have quicker access to the information. Some vendors have technology to help you de-duplicate information and move select data into online or near-line archives to enable search and rule-driven retention and destruction.
In summary, tape not only remains relevant, but is an important tool for organizations to recover from failures and, when used appropriately, to retain information over long periods in a cost-effective way. While disk-based backup offers great promise for many organizations, today it is primarily being used as a complement to tape backup. As an online media, it provides faster access and the ability to minimize storage redundancy. However, it doesn't offer the same low-cost, long-term storage characteristics that tape offers, and it doesn't offer the same isolation from disasters.
It is time for organizations to re-think their plans to eliminate tape. Many have confused the culprit. The goal instead should be to improve their backup and archiving processes. For most organizations, these improved processes will leverage the online nature of disk and the off-line nature of tape. Combined, they provide the best combination of performance and cost.
Protecting Your Data: It's Not Your Father's Encryption
Data Loss Prevention: Where Do We Go From Here?
Geoff Nesnow, a 12 year IT industry veteran, was a key player in the development of the LiveVault online backup solution. He joined Iron Mountain in late 2005 when Iron Mountain acquired LiveVault. At Iron Mountain, Geoff focuses on bringing new technologies and products to market. His experience with tape backup started in his role as IT consultant prior to starting with LiveVault, and he has written a number of published articles about storage and backup.