To say that tape is currently viewed as a strategic initiative in most organizations could at best be described as optimistic and at worst a fabrication. But the continuing growth of rich media (social media in particular) and unstructured file data, much of which appears to be destined for the cloud, are creating an unprecedented demand for economical back end storage on which to store it. Tape is now better poised to become that storage media of choice but it still has a lot of growing up to do in order to gain broad market acceptance.
This new demand for economical storage is what tape providers are counting on to fuel their future growth and stimulate a rebirth of the tape industry as a whole. While the last 10 years cannot be viewed as a lost decade for tape (after all, LTO went from its initial release of LTO-1 in 2000 to holding about 87% market share by the end of 2010,) it has suffered a perception problem in the eyes of end users. Many no longer view tape as being a viable media for data storage because of past issues regarding its durability, management overhead and difficulty in easily accessing data stored on it.
But organizations really need to look at tape anew for the following three reasons.
Tape’s new use case is archive, not backup. I do not know if anyone will ever convince me that tape technology will ever sufficiently change that it is again appropriated for use as a primary target for backup. In my mind, that ship has sailed.
However when used as an archive, it is storing production data that has a low probability of needing to be accessed again at some point in the future. The twist here is that when used in the context of archive is that tape cartridges will permanently reside in a tape library and do not have to be continuously handled and moved as they were when used for backup.
Tape is more durable. The saying used to be regarding tape cartridges that if you “drop it, throw it.” That is no longer the case. Granted, tape cartridges should never be confused as replacements for hacky sacks but their construction has sufficiently improved to the point where the integrity of the data they hold should not be questioned as a result of routine handling.
Tape has broken its allegiance with backup software. The introduction of LTFS (Linear Tape File System) means that eventually any application and/or operating system can write to or read from any LTO tape library as easily as they write to and read from disk. While the LTFS standard is arguably still in its infancy (about 1 year old now) and tape vendors have to police themselves so as not make the implementation of LTFS proprietary, LTFS makes it much easier to access and retrieve data from tape as it eliminates the need to first have to go through backup software to do so.
So by changing its primary use case, improving the quality of its products and standing on its own (no longer tied to backup software,) tape has opened itself back up for inclusion in many more conversations about how it can be seen as a viable storage alternative for this burgeoning “Big Data” onslaught.
But BIG questions still remain that the tape industry has to answer in order to overcome the built-in and justifiable resistance that organizations are bound to have to their broader adoption for tape in this new role.
Tape still reads data linearly. Tape is still tape and there is no way (at least that I am aware of) where it will ever be faster than disk when it comes to quickly accessing random files stored on a tape. That said, tape has come a long way in improving seek and access times for files located almost anywhere on a tape cartridge as a file located on a LTO-5 tape cartridge mounted in a tape drive can retrieved in 60 – 75 seconds.
But that is still a helluva long time for a user to wait if they do not understand that the file is stored on tape and the wait time could be as long as five minutes if the tape cartridge has to be retrieved from somewhere in the tape library and mounted. Further, I don’t see those wait times as being “generally acceptable” especially for users accessing social media data over the Internet who are accustomed to instantaneous responses.
Feeding data fast enough to tape and then reading it back again. Current LTO-5 tape drive speeds are rated at 140 MB/second native and 280 MB/sec compressed. Those speeds are great assuming you can implement disk systems and build network pipes that are fast enough to feed data at these rates to tape or accept data from it. But if servers cannot do that, shoe-shining occurs which slows both read and write rates and decreases the life of the tape cartridges and drives.
Ability to service multiple simultaneous read requests of different files. This question was never asked while I was in attendance in the Tape Summit but since tape reads data linearly, how will it handle multiple simultaneous read requests for different files residing on a single tape cartridge. If it queues these requests up and the files are spread out on the tape cartridge, you might be able to set your alarm and take a short nap before you get your file back.
Now I do not want to conclude or sound overly negative about tape’s prospects going forward in its new role. One of the more exciting conversations that I had while at the Tape Summit was with Rob Sims, the CEO of Crossroads Systems. Crossroads is working on a new product (Strongbox) that is designed to address and overcome many of these issues associated with implementing and using tape in this new role as a storage target for archive data.
But by Sim’s own admission, this new product is at best scheduled for beta release this summer and general availability in Q4 2011. So from an enterprise perspective, even if it does everything that Sims says it will do in regards to addressing these issues with tape (and he is promising a lot,) it is still two years and probably four years out before most enterprises would seriously consider using its Strongbox technology.
The Tape Summit that I attended the last couple of days in Las Vegas convinced me that tape is back in the storage conversation as a relevant and strategic storage media for solving some of the issues that storing Big Data creates. But it also convinced me that tape has a ways to go before it overcomes the prejudices and built-in bias that many users have developed against it and it may take a number of years and a lot of education and innovation before tape regains the stature and importance it once had in the data storage industry.