Disk-to-Disk-to-Tape: Four Top-Notch Bundles

Cobbling together solutions for disk-to-disk-to-tape may actually drive the cost of Tier 2 disk above the price of solution sets that combine the functionality.

A friend of mine, who is also CTO for a leading storage vendor, likes to tell a story about a recent sales call. An amusing (and probably apocryphal) tale, I repeat it here only to help frame a broader discussion of Disk-to-Disk-to-Tape (DDT) platforms and their potential role in both large enterprises and small-to-medium-sized businesses (SMBs).

Seems that a sales person representing the vendor in New York was having some difficulty selling the company’s DDT platform to a prospect and asked the CTO to intervene.

“This customer, who was looking at our enhanced backup platform, now says that he has decided to go with ‘Jay-BOD’ instead,” the sales guy lamented. “Can you maybe talk to him about it or something?”

My friend agreed and phoned the prospect, “Hi, I’m [John Smith], CTO of XYZ Corporation, and I understand that you are evaluating our enhanced backup product.”

With a thick-tongued Brooklyn lilt, the customer responded “Well, thank you for calling, but I am afraid that we have decided to go with JBOD…”

My friend retorted, “Well, you have good reason to be afraid…”

“Are you threatening me?” the customer snapped back, to which my friend quickly clarified, “No, it is a common misperception that CTO means Chief Threat Officer. However, I am the Chief Technology Officer, and I think there are several good technical reasons for you to be afraid about making the choice for Just a Bunch of Disk (JBODs) versus our platform.”

Well, to hear my friend tell it, complete with “wise guy” gestures and accents, the story is side-splittingly funny. Truth be told, however, there are some good reasons to go with Disk-to-Disk-to-Tape, as many companies are discovering.

DDT provides a way to expedite backup by first copying data directly from one tier of disks to another, then enabling tape backups as a separate operation without concern about backup windows. The second tier of disk (whether provided by JBODs or some enhanced disk platform) can serve both as a staging area for the backup—a cache, if you will—where useful data preparation can be performed, and also as a “nearline” repository for data that can be used to expedite discrete file or data block restores.

As a pre-tape backup “staging area,” the functionality of Tier 2 disk may be augmented with technology like Avamar Technology’s commonality factoring, which can be used to reduce the volume of backup data to a fraction of its original size. Data hosted on Tier 2 can also be expunged of “contraband files” like MP3s or AVIs, checked for virus signatures, compressed, and/or encrypted—all prior to backup. And, when open data naming schemes are fully developed, Tier 2 may also serve as a location for placing descriptive headers onto files so they can be migrated to appropriate tape resources such as Write-Once Read Many (WORM) volumes or to optical disk based on data preservation and retention requirements.

All in all, there are good technical reasons for deploying DDT. The industry knows it and has responded with a variety of proprietary DDT platforms at the high end, and a plethora of inexpensive JBOD solutions at the low end of the product line. Sorting through the options can be a dizzying experience, especially in the absence of meaningful head-to-head solution comparisons.

My friend’s argument that simple JBOD-based solutions may not be the best approach has some merit. Tier 2 disk, after all, is serving as a surrogate repository for mission-critical data. To perform this role even minimally well, some sort of RAID is probably required as a stopgap against the traditional foibles of disk.

RAID implementations range from software-based to hardware-assisted approaches (e.g., solutions from Promise Technology and certain Adaptec products), to full-out intelligent controller-based approaches (like Intel’s well-respected high-end RAID controllers). All RAIDs are not created equal, and you pretty much get what you pay for.

Augmenting Tier 2 JBOD with RAID capability is an absolute minimum configuration requirement for DDT. Additionally, you will find that some investment will need to be made in software to derive real value from the approach. You may need NDMP support for back-end Tier 2-to-tape interconnects, or something more exotic if you are writing to WORM or optical. And, of course, in addition to all the file scanning, data naming, and other useful purposes to which you may wish to harness the solution, you will need to purchase extra secret sauce technology.

In the final analysis, cobbling all of these features and functions together may actually drive the cost of Tier 2 disk above the price of solution sets that actually bundle the functionality. Four top-notch bundles that I am familiar with are Quantum’s DX-100, Avamar Technologies’ AXION, Arkivio’s auto-stor, and the soon to ship iStoRA from Breece Hill.

DX100 is a nice improvement on the groundbreaking DX30 platform released by Quantum a couple of years ago. DX30 was essentially a tape target surrogate, designed to drop into existing tape backup environments and add speeds and feeds to the laborious process of backup. DX100 retains this disk-as-tape paradigm (for now) but improves on the capacity (up to 64 TB of disk), performance and pricing of its solution. Wedded to its MAKO PX720 Tape libraries, DX100 delivers as much bang for the buck as large-scale tape backup shops may ever need.

By contrast, Avamar’s AXION and Arkivo’s auto-stor are more than tape emulation targets in a DDT configuration. Each solution brings a “data lifecycle management in a box” play to the market that challenges the depth and capability of EMC’s primordial offering in this space: Centera.

AXION, which comes in two flavors—“E” for branch offices and smaller settings (1.2 TB of usable disk per unit) and “M” for larger enterprise settings (scalable to exabytes of storage, according to the vendor)—migrates data after eliminating redundant data sequences, enabling companies to derive better capacity utilitization from their disk investments. The technology also takes a page from EMC and provides content addressable storage to track data location as it migrates from platform to platform. Currently, the solution ships with all software and hardware supplied by the vendor, though it is conceivable for CEO Kevin Daly to eventually separate the two so that other hardware platforms can be supported.

Arkivio’s solution provides some interesting functionality for policy and access-frequency-based data migration, so that your less frequently accessed data doesn’t live on your most expensive gear. The technology is being wedded to various low-cost disk array platforms, like Nexsan’s ATA-based ATAboy2 arrays, to facilitate a one-stop shop for hardware and software.

Meanwhile, Breece Hill, under the guidance of CEO Phil Pascarelli, is about to go to market with one of the first of a burgeoning class of disk/tape hybrid products designed to give backup relief to the SMB market: a product currently called iStoRA. At a meeting with Pascarelli’s team in Louisville, CO recently, I was impressed by the combination disk array (1 TB of Serial ATA RAID) and half-inch tape autoloader (2 TB raw capacity)—and by the value-add supplied by Avail Solutions Integrity software, which has been integrated with the box. At its anticipated price point of less than $10K, this appliance holds the potential for setting the bar in drop-in, policy-based, DDT solutions for the SMB market.

Based on conversations with other tape vendors at the Fall Storage Networking World conference in Orlando, there are a lot more products on the horizon to meet the needs of this space. If you have any input to offer on DDT, or experiences—positive or negative—with these products, drop me a line at jtoigo@intnet.net.

About the Author

Jon William Toigo is chairman of The Data Management Institute, the CEO of data management consulting and research firm Toigo Partners International, as well as a contributing editor to Enterprise Systems and its Storage Strategies columnist. Mr. Toigo is the author of 14 books, including Disaster Recovery Planning, 3rd Edition, and The Holy Grail of Network Storage Management, both from Prentice Hall.