Tape Can Still Close The Backup Window
December 12, 2011 Dan Burger
Every aspect of your business is bombarded by the message that modernization is the key to happiness. IT is just one piece of the puzzle. And even that puzzle has dozens of pieces that require decisions on whether to replace or refurbish. Replacing may result in the biggest benefits, but often there are no guarantees, particularly if what’s in place is working. Finding what’s not working is the first step and then deciding if you can improve what you have without replacing investments in hardware and infrastructure is the next step. Let’s take storage for example.
Disk-based backup is one of those raging wildfires of technology that pits modern products against legacy products. There are benefits and risks to both sides. However, when you consider the large investments in hardware and infrastructure required with some of those choices, you’ll find that most businesses in the small to midsize category are not eager to jump into a pool until they assess how deep it is. Total cost of ownership is a serious consideration and it cuts both ways. Those that move may pay for wholesale infrastructure changes. Those that stay may have maintenance costs that nickel and dime–or hundred dollar–them to death.
The reality in the IBM i storage picture is that tape remains the key ingredient in most companies’ storage strategies. Even when you look beyond the IBM i community you find that around 75 percent of all data is stored on tape, so skip the IT Neanderthal comparisons, or at least don’t confine them to the IBM midrange shops. Why is tape still dominant? Because tape has proved to be a trusted and cost-effective component for archiving and backup, says Ed Ahl, director of business development at Tributary Systems, a developer of storage virtualization software for IBM i and z/OS systems.
So, a track record does count for something in this buzzword-of-the-day business.
Replacing tape with disks is an alternative that has growing interest, and a lot of marketing and media (pun intended) support, but working with tape is the chosen media for the majority of companies here and now. Unilaterally referring to tape-based companies as laggards is disregarding individual company circumstances and in-place systems (call them legacy, if you like) that are doing well, thank you very much. It’s also being ignorant of the numerous improvements to tape that have taken taped-based systems to higher levels of performance and efficiency than many people (those who quickly dismiss tape) realize.
Admittedly, there are companies that are woefully behind the times. Their neglect to move forward, even while maintaining an adherence to tape, earns them the laggard tag. And they become the poster children for what’s wrong with tape.
But my point today is that tape is not being overshadowed by disk and the new kid on the block, flash storage, is still wet behind the ears at this point, despite having a great press agent. Not in the IT departments of the average IBM i shop anyway. Ahl is a great believer in keeping things simple and inexpensive.
“There should be no penalty for not moving at the pace created by forces outside of your own company. If the current functionality works and there are alternatives that allow you to improve what you have without making a big jump, an incremental approach makes a lot of sense,” he says. “Better, more secure backups today with the flexibility to move to new technology when the situation is right should be a choice not a mandate.”
Keep control of your own environment is sound advice.
All I’m saying is that products are available that can help existing tape-based systems improve without jumping to new technologies, disrupting business processes, or possibly breaking the bank. If you can legitimately determine that eliminating tape-based backup is your highest IT priority, then by all means explore the alternatives that better fit with your IT roadmap. But if your IT budget has to accommodate other issues with equal or higher priorities–application or database modernization are two possible examples that come to mind–know that you still have options to improve what you have.
To my way of thinking, this is the reality of the midmarket. It’s not solely a reflection of the current economy. That’s the way the midmarket has always operated. Stick with what works, enhance it to improve business efficiency, and turn your attention to things that don’t work. Spend your IT budget there.
In my chat with Ahl last week, he agreed with me that the most overused sales lever in any discussion about storage is performance. Everyone who sells storage wants to talk about performance. Everybody that buys storage wants to talk about performance, too. However, the sellers usually want to talk about it before the implementation phase. The buyers often want to bring it up again afterward. Performance figures vary wildly when comparing sales literature with real world situations. Even when a large dose of overpromising is being administered and it’s duly noted and taken into account, high expectation fever can cause temporary blindness.
Most companies are vulnerable to this fever because there is immense pressure to do something about backup completion times. Few things motivate storage purchases like a bottleneck in data backup causing longer than expected backups and keep the IBM i from devoting its considerable heft to doing actual work. The promise of pain relief has led to many quick decisions.
“Most midrange shops who have been through this will tell you the new stuff that is coming out doesn’t fit in their environments. If the answer is to put the new storage hardware or software on a new Power7 system, or to spend a considerable amount on infrastructure improvements, that’s not a good fit for everyone,” Ahl says from repeated experiences.
One example he uses is in shops where modern LTO tapes are being used to beat the clock on tight backup windows.
“Data deduplication devices are not a means of shrinking backup windows in these shops,” Ahl says, “even though it is widely assumed that it will do the job. They could give you more storage for less money. They can crunch a lot of data. But you give up performance in environments where more modern LTO tapes are being used.”
Speed is alluring, but there are major concerns about data integrity as well. This issue relates back to staying with a known system. It’s a comfort zone and some would criticize the reliance on a tape-based security blanket as a sure way to miss the boat on modern technology. I think that’s a false assumption and Ahl agrees.
“Companies want to backup everything from the OS to data and be able to restore their iSeries straight out of a physical tape or a via a disk layer. They want to maintain their current environment without so much sacrifice and be able to isolate the physical layer so that if any problems occur it doesn’t affect the box, which has already been released from doing backup and is once again doing production work,” he says.
Some of the problems that can arise with tape-based systems are in instances where a company wants to replicate data to a secondary location and create tapes there. LTO options can be limited in that regard and the LAN connection can be slow and tie up the production box. There is software to help with this and Tributary Systems is in the business of selling it. The software emulates an LTO tape library, writes to the disk cache of the underlying appliance, which generally allows backups to be completed much faster than disk. Tapes can then be created from the disk.
It’s one of the ways to get more out of a tape-based system without forcing a change in the basic way backup is done. Technology that doesn’t force change, allows substantial enhancement, doesn’t create dead-ends for applying other technologies. And it keeps things simple.