Moving from “D2D2T” to “D2D2D” is sometimes seen as an unattainable hurdle that enterprise organizations cannot overcome when tape is used for secondary roles such as archiving or disaster recovery (DR). But replacing the “T” in “D2D2T” with a “D” is now practical, possible and affordable. Doing so simply means enterprises need to demonstrate that disk offers the same or more functionality than tape when used in these capacities while costing the same or less.
Moving from tape to disk as a primary backup target is becoming almost a no-brainer. The obvious benefits that the majority of enterprises realize are shortened backup and restore windows, near 100% success rates and the cost per GB of disk being on par with that of tape when deduplication is used. Equally important, IT Directors report increased productivity from their IT staff and improved control over their backup-related IT expenditures once backup to disk is implemented.
It is, however, the use of disk in lieu of tape as these secondary roles that requires more thought, planning and rationale on the part of enterprises in order to execute on. In this secondary role, disk must now assume some of the other strengths that are historically associated with tape such as:
- Portability. Provide a means to get data offsite to do disaster recoveries.
- Power and space efficient. Minimize ongoing “hidden” operational data center costs such as heating and cooling while not taking up precious data center floor space.
- Immutability. Ensure that data has not changed or been tampered with to satisfy data preservation requirements issued by federal agencies like HIPAA or the SEC.
- Long life. Confidence that the media on which the data resides is long lasting and that the data will be accessible in case it needs to be recalled.
So as part of an enterprise building its business case to move from “D2D2T” to “D2D2D” for archiving and DR, it needs to first show how disk can do as good or better job in delivering on these attributes and then doing them in such a way that disk’s costs and overhead are the same or less than tape.
One of the most common reasons to move disk to tape is to get the data offsite for disaster recoveries. But to a large degree issues around data portability have been largely resolved with the availability of deduplication and replication on today’s disk storage systems.
Using deduplication and replication to do these tasks comes at a price but so does using tape. The soft costs associated with tape include people loading and unloading the tapes, software that tracks the physical location and movement of tapes, vehicles to move the tapes and an offsite facility to store the tapes.
Tape also carries with it the added risk of losing cartridges plus the time and difficulty associated with performing restores. This risk associated with lost tape cartridges can be particularly acute from a legal and even a financial perspective as tape loss can lead to organizations facing uncomfortable questions about the integrity of their data.
It is for reasons like these that organizations such as the State of Alaska are already using disk in these secondary roles. To efficiently move its data offsite, it is leveraging EMC Data Domain’s deduplication to first minimize the amount of data stored and then replicating deduplicated data to minimize the amount of data sent over its WAN links. In so doing, it has eliminated the risks and costs associated with tape while improving its ability to archive data and recover from disasters to a level it could have never before achieved using tape.
The State of Alaska’s use of deduplication has had the additional effect of making EMC Data Domain systems as or more power and space efficient than tape. While tape proponents rightly claim that tape cartridges do not consume any power while sitting on a shelf, tape cartridges still consume energy when they are transported and even when they are stored as they still need to be stored in climate controlled environments. So to say tape cartridges do not consume energy or data center space is a bit misleading.
In this respect, the State of Alaska as well as a health care provider on the West Coast with which I recently spoke both share stories as to how their use of the combined deduplication and replication features on EMC Data Domain have minimized their power and storage consumption.
Case in point, the West Coast health care provider related that since it started using disk in lieu of tape as a secondary target, it has almost completed eliminated the need for its IT staff to drive to work to do recoveries. Further, it no longer needs to power up tape libraries in which tapes are stored since it can now store all of its data on EMC Data Domain systems at each of its six sites.
The ability to preserve the original condition of data on tape, or WORM (Write Once Read Many) capabilities, is also sometimes cited as a reason to use tape in order to satisfy specific legal requirements. However deduplicating disk systems like those from EMC Data Domain now possess this type of functionality. This comes in particularly handy for end of month full backup copies that enterprises neither want the data expired nor changed. In this circumstance, disk actually has an advantage over tape since if the data ever does need to be recalled, it can be accessed more quickly than when it resides on tape.
The final argument in favor of using tape over disk is tape’s long life which is rated at 30 years. But enterprises need to ask, “Is this feature a benefit or has it become a potential liability?” It can be forcefully argued that keeping data on the same media may actually become a detriment as it lulls enterprises into a false sense of security that their data is recoverable when the technology needed to recover it may have been obsolete for decades.
Consider this example. It is now 2010. Even if an enterprise has a tape cartridge from 1980 that is still viable, how is the enterprise going to recover the data from it? It minimally needs to put in place the right tape drive, the right backup software and then possibly even the right hardware and OS before it can even think about restoring any data. Then assuming it can even find the 20 – 30 year old technology needed to do the restore, the cost and time required to perform this task may preclude it from ever occurring.
Disk keeps enterprises from being lulled into this false sense of security as it gives them a simple and easy method to move data from one generation of disk technology to another. For instance, using EMC Data Domain’s replication technology, when a new solution is purchased, existing data can simply be copied from the old one to the new.
This technique avoids technology obsolescence while enabling enterprises to take advantage of increasing disk storage capacities as it is fairly safe to say an EMC Data Domain solution 5 years from now will have approximately 8x the capacity and performance as a current solution.
Portability, power and space savings, immutability and long life are the four reasons most often cited for using tape in lieu of disk in the “D2D2T” equation. But once enterprises take a serious look at how well positioned disk now is to replace the “T” in “D2D2T” with a “D”, going completely to a “D2D2D” implementation is more than just a sound approach to effectively resolving their archiving and DR challenges. It also becomes the right choice from a business perspective.