By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Disk-to-disk (D2D), and disk-to-disk-to-tape (D2D2T) have taken the center stage in industry discussions around data protection. Both technologies leverage a multi-tier architecture fielded either with the intention of improving the performance of backups in production environments, or to expediting restores in the wake of a primary storage failure or accidental file deletion.
Data backup versus data copy
The only way to protect data, of course, is to make a copy. D2D and D2D2T are both topologies intended to enable more efficient creation and use of data copies. However, data copy and data backup are not synonymous.
In a backup, threads of data are joined together to form a backup stream, and possibly subjected to compression before they traverse an interconnect to a target storage device. There are many different interpretations of the term backup and a considerable amount of investigation may be required to understand what a vendor means when using the term.
Most file system backups, for example, actually entail two discrete data movements: one, as described above, comprises the stream of copied data itself; the other is usually a transfer of metadata (data about data) to an entirely different target. Files can be restored on a granular basis by selecting the metadata listing for the file then waiting for the backup/restore software to locate its corresponding data in the backup data set.
A variation on this theme is the "bare metal backup." Bare metal backups transfer a bit image of a volume containing data to a target and provide "all or nothing" restore of the data. In other words, discrete files usually cannot be restored individually.
Using a disk as a target for a backup process can yield speed advantages over tape. Backups can be done faster. Some D2D2T configurations place an intermediary tier of disk between the production disk and the tape device, where it serves as a "surrogate tape" device (that is, it pretends to be a tape drive). Data may subsequently be moved from this disk to the tape device itself as a separate process that does not interfere with production processes.
By contrast, data copy is simply the movement of a copy of data, with its metadata intact, from one disk device to another. Used in a data copy role, the secondary tier of disk in a D2D2T topology is treated not as a surrogate tape device, but as a disk mirror -- a repository for synchronous or asynchronous file copies. This doesn't mean that a conventional tape backup can't be performed between tier 2 disk and tier 3 tape, but consumers need to remember that copying files between tier 1 and tier 2 disks takes considerably more time streaming data between disks.
Bottom line: D2D2T topology says nothing about the way in which the intermediary tier of disks is being used. To understand the benefits and the drawbacks of D2D2T, you need to first understand how an application will use the topology, in other words whether the tier 2 disk will act as a surrogate tape target or as a mirror. The former may speed up backups, but does not enhance restore speeds. The latter may actually increase the time required to copy and transfer data, but can substantially reduce the time to data restore in the wake of a data disaster.
Some productized models
The simple distinction above speaks to differences in applications and is usually illustrated by vendors with three discrete storage devices and lots of arrows showing flow. As in the case of SANs, this can leave one with the impression that DDT requires multiple storage devices arranged in a fabric or other interconnect topology. Truth is that, while some vendors have offered DDT "solutions" that make use of this multi-box approach, appliances are also available that combine software with hardware in a single component: D2D2T in a box.
One that our lab has been testing is the Breece Hill iStoRA appliance. Breece Hill combines SATA disk (the low end unit offers about 1.5 TB of disk capacity) and a 10 tape autoloader. With resident Integrity software, from newly acquired Avail Technologies, Breece Hill has created a one-stop-shop data protection solution ideally suited to small and medium-sized businesses.
Breece Hill's box can connect to Fibre Channel, iSCSI, or directly to a corporate LAN. The Integrity software is deployed both to the appliance and to servers (including Windows, Linux and Unix brands) and desktops where server administrators and users can manage copy or archiving events. The disk and tape media on the appliance are configured into "vaults," enabling that data from different clients can be segregated and maintained. Software on the appliance periodically migrates data from the disk vaults into corresponding tape vaults as a separate off-LAN operation.
In truth, Breece Hill could use its disk as to emulate tape -- a target for the myriad tape backup software products in the market today. The company has decided, however, to add value to commodity disk and tape with Integrity software, which delivers the additional capability to support hierarchical storage management and archive: the primordial building blocks of information lifecycle management.
Just on the other side of Denver, Spectra Logic will shortly be shipping its RXT platform. Using its robotics technology developed for tape automation, the company has innovated a removable disk subsystem that takes SATA drives and "picks" them with the same robot used to mount and dismount tape media.
Unlike the Breece Hill approach, which arguably sacrifices backup speed in exchange for fast restore speed (and for improved data management generally), Spectra Logic prefers, in its initial purposing of RXT, at least, to stick to conventional backup modalities. They use their removable disk to emulate tape, thereby enhancing backup speed by writing streaming data to removable disk. Discrete file restore speeds, however, remain as slow as they are in conventional tape.
With the right software (maybe Integrity), Spectra Logic may eventually be able to add additional value to RXT, bringing the simplicity of the Breece Hill approach to enterprise customers. For now, the important thing to keep in mind is that the term D2D2T does not require the implementation of multiple storage arrays. It may in fact be a one stop shop.
For more information:
Backup School: Lesson 2 -- Which backup media is right for you?
About the author: Jon William Toigo has authored hundreds of articles on storage and technology, including his monthly backup/recovery feature for SearchStorage.com. He is also a frequent site contributor on the subjects of storage management, disaster recovery and enterprise storage. Toigo has authored a number of storage books, including Disaster recovery planning: Preparing for the unthinkable, 3/e. For detailed information on the nine parts of a full-fledged DR plan, see Jon's Web site.