How are tape technologies changing today?
From a media standpoint, we are poised to see a huge capacity increase in tape cartridges -- both the proprietary “enterprise” media and the generic Linear Tape Open (LTO) standard cassettes. Let me contextualize what I mean by indulging in a bit of historical retrospective.
For many decades, tape technologies have advanced through the innovation of the media and automation folks within the tape industry, and also by borrowing technologies from the disk world. Vendors like IBM and FujiFilm have done a lot on their own to advance tape technologies, including the improvements we have seen to servo track systems over the past decade that have dramatically reduced head alignment-related errors. Gartner's statement that one in 10 backup tapes [will] fail on restore was an exaggeration when they first said it in 1999, and is completely out of step with present reality.
The latest set of improvements in tape media have been associated closely with technology borrowed from the disk world, like giant magnetoresistive (GMR) read/write heads and partial response maximum likelihood (PRML) algorithms. That combination of technologies lets us detect weaker signals stored as smaller bits, and has accounted for improvements in media track densities and media capacities, first, in the disk world, [and] then in the tape world.
When disk makers introduced perpendicular magnetic recording (PMR) -- the technology that enables 2.5- and 3.5-inch SATA drives to present capacities in excess of a terabyte today as a function of making small bit magnetic states perpendicular to the recording surface, rather than running in parallel to the media as was the case in the past -- the tape doomsayers again came out in force.
They claimed that because tape lacks a rigid media surface, it couldn’t leverage PMR and would be left behind as disk capacities grew.
Then, in 2010, FujiFilm and IBM demonstrated something called called “perpendicular orientation” or PMR Type II -- leveraging a new process for coating tape with Barium Ferrite (BaFe) particulate.
BaFe has some cool properties, including the way that it stands magnetic fields up, again, perpendicular to the media surface, thereby enabling a huge capacity bump in the media without requiring significant changes to recording heads in the drive. IBM demonstrated recording densities of 29.5 billion bits per square inch with LTO standard-size tape cassettes, so we can expect to see 35 TB uncompressed cartridges in the near future.
Tape drives and automation features are also getting better -- look at the brainy management system implemented by Spectra Logic on its rigs, for example. Plus, add-in appliances from tape industry ecosystem partners like Crossroads Systems’ Read/Verify Appliance and Strongbox are building out a resiliency and reliability story for tape that beats what we have seen in the disk world.
Bottom line: Tape will shortly provide a low-cost, high-capacity alternative to disk for much more than data protection. Think active archiving, or even NAS on steroids: petabytes of storage on a single raised floor tile consuming the electrical equivalent of a couple of light bulbs. Show me a disk array that can deliver those power and capacity/density metrics. Bet you can’t.
This was first published in February 2012