Based around some recent tape industry activity, I recently blogged about the fact that tape is more than alive and well ... in certain circumstances and for the right uses it actually merits more credit and attention than it has typically had. I finished the piece thus: "Perhaps the biggest obstacle the tape industry faces is indeed that not all decisions are made logically. Tape certainly has some "historical baggage" ... and today that manifests itself as a set of largely erroneous perceptions. It's the emotions that will matter for the next few years. Simple economics might force users to bite their tongues over the coming years--hating (or is that having?) to admit they love tape again." Whether it's due to serendipity or a scotoma, there seems to have been an uptick lately in group discussions and articles on the use and abuse of tape. I'll cover some of the key points here very briefly ... but the real point of this column is to reflect on the fact that such a supposedly fact- and detail-based business can be quite so emotionally based. It's like the storage crusades or primaries!
The summary situation is this: Because it is generally (and understandably) perceived as outdated, tape is a technology people love to hate. But the things that made it unlovable mostly belong to the past. Today, tape is arguably on the cutting edge in providing highly cost-effective and reliable answers to certain data storage challenges. It turns out that what is outdated is not tape itself but the perceptions of tape. While we've witnessed and embraced the rapid evolution of--say--processing power and solid state memory, a similar and no-less amazing forward leap in storage capacity, searchability, and reliability has been occurring in the tape world. Multi-TB cartridges are commonplace, and the enterprise tape business (still driven by some of the biggest IT names--IBM, Oracle, HP, and others) represents something in the region of $2.5 billion annual revenue for drives, libraries, and media.
Of course, there have been legitimate historical reasons to avoid the technology or to try to reduce its use, but the main ones have been, or are being, addressed; colloquially stated these reasons, and the quick ripostes, are:
-- "I can't find my data": There are many file systems that now work with tape and the LTFS tool makes tape cartridges "self-aware"--in other words, they are properly portable between systems with their contents shown upon being mounted; basically making tape cartridges just like a giant USB thumb/stick drive.
-- "Is the data actually there and still OK?": There is an increasing number of tape analytics tools that monitor and check the health of everything-tape: is the data actually written, can it be read, and then verifying that at user-determined schedules.
-- "Yeah, but c'mon, tape just hands-down isn't very reliable, is it!?": With modern drives and media, this is simply not true. The raw reliability of the media (measured in terms of hard errors) is in fact orders of magnitude better than it is for disks (why do you think we have RAID!?). To put this in perspective, for a mid-sized environment, disks would typically have a hard failure every few days or weeks (depending whether you're talking SATA, SAS, or FC), whereas tapes might be months (LTO) or many years (enterprise drives) between equivalent failures. Who knew, eh?