Disc or Tape: When, Where and How
In the headline above, I’ve set up disc and tape methods of storage in a simple dichotomy. However, you storage networking pros out there know it’s not always a choice between one and the other. I’d guess most of you work at places that use both, and you know how disc and tape solutions support each other in an overarching storage schema. It’s not really about which one’s better. Both discs and tapes have their roles to play, and both do their roles well. What we’ll be looking at here is when and where discs are better for data preservation than tapes and vice versa.
Most large organizations use discs as their principal means of storage, aided by massive configurations of refrigerator-sized disc-drive boxes. The majority of large organization also use tape to record and recover information in the event of a disaster. “Discs are your primary storage,” said Wayne Adams, chair of the Storage Networking Industry Association (SNIA), who also works for EMC. “Tape tends to be the final, long-term archival medium. If I own an oil refinery, the typical lifecycle of that is 50 to 75 years. If the plant has all this operational and maintenance data, 30 years from now, when you reference that history, you’re going to want that snapshot in time in 2006.”
Storage wasn’t always so complicated nor did it take such a long view. However, the increase of regulatory compliance, data production and data mobility has created a demand for sophisticated storage systems, Adams said.
“Years ago, the relationship was called direct attached storage, where one or more disc drives were allocated to a single computer. As there’s been a proliferation of the need for information to be shared across multiple computers, (it) has mandated certain economies of scale and efficiencies,” Adams said. “In the very broadest sense, our whole society is immersed in an information explosion. You have it in your home and at your desktop at work. Businesses have it in customer-relationship data and regulation data. With that going on, there’s a need to figure out what type of data is needed and at what speed and access time.”
To determine how an organization should save data, storage professionals should consider how rapidly the organization can recover lost data. Where speed is concerned, disc is the superior storage tool. “If you have it on a second disc that’s spinning right there, then your time element is quick,” Adams said. “If it’s on a tape drive that was archived five years ago, then it might be anywhere from a couple of hours to a couple of days based on the service level agreement you have with the company that stored that away for you.”
Lower-capacity discs are quicker and more useful in time-sensitive information transfers, which are common in banks and other financial institutions. “In the disc-drive space, you have trade-offs in capacity and speed,” Adams said. “Typically, high speed will have lower capacity because there’s a mechanical arm that goes across the platter. If I had the choice of storing a 100-gigabyte disc drive and 300-gigabyte disc drive, the latter means that there’s more travel for that arm because you have more data on the platter. The high-speed, low-capacity use of a disc drive is for very transactional types of data.”
Slower-paced institutions with a substantial amount of data to store—for example, a library and its voluminous books—will use discs, but they’re usually low-speed, high-capacity discs, Adams said. “The value of the data from their catalog isn’t the equivalent of a bank transaction. They’re going to be using lower-speed disc drives with higher capacity because it will cost less and they can store more.”
In addition to speed, storage professionals should consider how far back into the records they’ll need to go if disaster strikes. “If you’re in the midst of running your business today and you have a failure of some kind, how far back do you go to reset your business?” Adams said. “Can you go back to one second, or can you tolerate going back to the start of the day? Everyone thinks they would go back to one second. There are certain applications like personal e-mail, where if you miss two or three e-mails for the day, you can just tell them to resend it. But if that was a banking transaction that was moving seven or eight digits worth of value between businesses, you can’t afford to lose that transaction. You’ve got to go back to the start of that transaction and be within a fraction of a second to that recovery point.”
Tape has an advantage here because it lasts longer, but discs have longevity too. “They’re both rather stable—they last a long time,” Adams said. “In the case of tape, you end up with a cartridge format. You get a longer shelf-life with tape than you would over a disc.” Many tapes, unlike discs, also have protective physical surroundings. Adams said archiving company Iron Mountain stores many of its clients’ tapes in an underground facility in the Rockies. “There’s always what I call the ‘final resting point.’ You want to put it in the most safe, secure location.”
–Brian Summerfield, email@example.com