When the question of data archiving gets raised, tape devices usually spring to mind. Quite a few generations have cut their teeth on computers that loaded and saved data to and from a tape drive -- once a humble audio cassette recorder.
Initially, tapes were far less expensive than any other form of archiving. Early hard drives and optical storage were so expensive they were only contemplated as primary on-line storage or (in the case of optical) a long term but expensive alternative to tape. Optical storage initially benifited from compatibility. There were only a relatively small number of standards for optical drives and the chances were your devices could read most of them -- whereas there was (and still is) a bewildering array of tape formats and standards.
Many people confuse archiving with backup. Archive data still needs to be backed up just like regular data. For backup the most logical choice is still tape because it allows users to store their backed-up data offsite. Disks are not yet very practical for doing this, unless fibre-attached over some distance in a metropolitan area network (MAN). Even then, a remote disk site within 10 kilometres or so could easily be affected by the same disaster that takes out the primary data centre.
While tapes remain a much safer bet for offsite/remote backup, the truth is that tape systems are not always the best or even the most convenient methods for archiving data.
Given that the cost of hard drives is plummeting, it is no surprise that the method of choice for many is a RAID. With IDE or even SATA drives, the cost of a beefy chunk of archival storage is becoming quite low.
This is not to say that there is just one best solution because there are numerous different levels of archiving. More important or younger data will need to be accessible instantly, while older, less important data might only be needed periodically -- enough to still warrant keeping it onsite, rather than in a data vault.
In the first instance, the speed would suit a RAID. The second where speed is not as important could be carried out by some of the faster tape-drive libraries.
There is, however, a third alternative bridging both worlds to provide a neat answer to the problem of data throughput with some tape drives -- VTL or Virtual Tape Libraries.
A VTL is simply a disk array that is capable of simulating tape devices so that it appears as a tape drive to your archiving or backup software.
Many vendors provide VTL solutions that utilise their own hardware and either their own proprietary VTL software solution or that of a third party vendor.
Software is potentially more crucial to archiving than the media itself, if you consider the potential need to access today's data in a usable format 10 years from now (think of how fast technology changes, 10 years ago 1GB was massive -- not to mention document and data structures). Archive searches must be quick, customisable, and accurate to ensure that the snapshot in time can be located as efficiently as possible. There is absolutely no point in having a warehouse full of boxes of paperwork from the last five years if there is no structure to it. An employee needing to find a single page of a memo could be looking for the proverbial needle, let alone following a paper trail of documents. The same goes for data that is archived; if it is not readily accessible, readable, and more importantly searchable then all can say is "good luck".
Oh, and in case you are thinking that archiving data only protects your bottom line by providing timely restoration and retrieval of business data, then think again. In Australia you must retain all your financial data in case the Australian Tax Office asks to take a peek at it, and if you deal with any US companies then there are quite a few compliance hoops you might need to jump through.
The majority of vendors participating in this review offered disk based products for evaluation. The disk specific products range from Network Attached Storage (NAS) to Direct Attached Storage (DAS) arrays -- usually combining some form of RAID. It becomes difficult to directly compare each of the devices, so for this review, each of the products is assessed in its own right.
