The New Economics of Data Archiving

What does a savvy business do with the mountains of data it creates? According to Chris Wood, CTO of Sun’s Data Storage Management Practice, the answer lies in an intelligent, tiered archive that leverages the economic advantages of tape. Sun Inner Circle recently sat down with Wood to find out more about why tape makes sense from both an economic and a competitive perspective.

Inner Circle: What’s driving the need for long-term data archiving?

Chris Wood: Essentially, three things are driving long-term archiving: Increased regulatory compliance requirements, the need for good corporate governance, and the realization that historical information can be a competitive weapon. For many organizations, these three things are doubling the amount of data that needs to be archived every year.

The potential cost savings of intelligently archiving information should not be overlooked. I know of an automobile manufacturer that now archives information about every step of the car manufacturing process. This reduces the number of cars that need to be recalled if components malfunction — the company simply identifies the facilities and machinery used to produce the cars with problems, and then recalls only those vehicles instead of the entire fleet out on the road. That’s just one example of how archived information can have a significant impact on the way an enterprise does business.

IC: How do you know when data is ready to be archived?

CW: Like most storage matters, archiving decisions need to be based on the cost of holding on to stored information. It’s important to be able to justify data needed on spinning disks. Most organizations struggle with how to improve systems they need for everyday operations, which often become clogged with infrequently accessed information.

It’s also important to understand the difference between backup and archiving. For example, an email saved on an email server backup tape is not archived data — it’s just a backup. Archiving, however, removes old data from the active database, which gives everyday information more room to function efficiently.

Moving data to a protected archive will not only improve the performance of the servers in the database, it will also substantially reduce backup requirements because there is simply less data to back up. But it can be tough for many companies to identify data in need of retirement, which is why Sun provides assessment services that help organizations analyze the appropriate data candidates for archiving.

IC: How can companies plan for future data access if they don’t know how file formats will change?

CW: Data migration costs money, so the job of a well-designed archive is to reduce migration to once every 20 years or so. An open storage format is critical for long-term savings because file formats are certainly going to change — and devices and applications that know how to read certain kinds of data today may not exist in the distant future. The worst thing an enterprise can do when archiving is to write data in a proprietary format.

Sun has historically allowed future generations of tape drives to read prior generations of tape media formats. In fact, today’s Sun StorageTek 9840 series can read the first generation of this tape library family. That’s close to 20 years’ worth of built-in backward compatibility.

IC: So what makes data archiving cost-effective?

CW: Cost-effective archiving largely revolves around old data being accessible in its original format, regardless of whether the data is stored on disk or tape. For one thing, this allows an organization to archive data in locations that make the most sense. But making historic information accessible through the original applications also removes the extra step of manually extracting data and then converting it to its original form. That way, a word processing document remains a word processing document and a spreadsheet remains a spreadsheet when you access it.

Sun takes care of this problem with a software layer called the Sun StorageTek Storage Archive Manager, or SAM, which presents a unified read/write archive storage interface for applications — and it doesn’t matter what kind of storage media and network is in place. SAM also allows users to set policies so the software layer can control where data is placed, which saves a great deal of system administrator time.

Plus, SAM is designed to protect existing data storage investments, because the software layer can restore any kind of data, no matter what kind of application originally wrote the data — or the kind of network in which the data is stored.

IC: You’ve mentioned tape a few times. Why should tape play such a large role in long-term archiving?

CW: Tape is extremely economical when it comes to power consumption, storage capacity, and physical real estate. It ends up being more cost-effective to keep everyday operational data on disk, and then place the rest in archives. Think about it: Tape at rest consumes no power, and if the data it holds is infrequently accessed, why waste energy on something that isn’t needed all the time?

Tape cartridges don’t require a raised-floor environment, which frees up valuable real estate for systems that really need the space in a datacenter, such as servers, rotating media, and networking gear. Plus, with a solution like SAM, historic information remains accessible and can be easily transferred to disk.

IC: How does tape’s storage capacity fit into the economics of archiving?

CW: Tape wins every time when it comes to the overall cost of holding data and reducing the amount of space needed for an archive. With the Sun SAM file system, tape can be filled to the brim in about the same amount of space as a disk — but it holds twice as much data. Remember, most disks have utilization rates in the neighborhood of 50 percent. What’s more, tape can compress data by a factor of two to one. So combined with the space savings provided by SAM, tape can potentially hold 400 percent more data than disk.

Besides, from an initial purchasing perspective, tape is a just a great value. Tape media in a well-maintained system will last three times longer than the average five-year life span of a disk drive. And with this longer life span, all the other activities that data requires — copying, moving, replicating, and migrating — cost less, because they don’t occur as frequently as they do with disks.

IC: What security issues are associated with long-term storage?

CW: An archive must be secure — otherwise it’s not worth the investment. That’s why built-in encryption is vital to archive and storage security. A number of Sun archiving solutions make the investment worthwhile by being fully encrypted with AES-256, the leading crypto algorithm approved by the National Security Agency (NSA). This includes tape drives like the Sun T10000 and the recently announced 9840D, as well as the Sun SL8500 and 3000 libraries. So with these solutions, if unauthorized persons get hold of an archive tape, it’s very unlikely that they will be able to access the information. Instead, the tape will simply be lost rather than the data being stolen.

Sun plans to introduce file system encryption throughout its storage offerings over the next few years, and in the Sun ZFS file system by the end of this year. This will allow data to be selectively encrypted on any media, including optical drives, holographic storage, and disk drives. For owners of data archives, selective encryption will mean that even in the event of a firewall breach, the intruder will not be able to read data.

IC: What else is in store for data archiving down the road?

CW: Data archiving, as I mentioned earlier, is fast becoming a competitive differentiator. But it needs to be implemented both economically and efficiently, and that requires the expertise to understand which media are most appropriate for long-term storage. Sun happens to offer the complete spectrum of storage systems — from the highest-performance disk and tape systems to the fastest data access on tape. I know of no other organization that can make this claim, and I’m certain this background will serve our customers well in an era where being thrifty is part of being operationally efficient.

About Chris Wood
Leighton C. “Chris” Wood Jr. is director and chief technology officer of Sun’s Data Management Storage Practice. He is responsible for identifying and delivering technological solutions that best address complex data management problems. Prior to being named to his current position in 2007, Wood was director of technical sales and marketing in Sun’s Global Network Storage Sales, where he oversaw technical support, education, and product development for network storage solutions.