LTFS (Linear Tape File System) was introduced just over two years ago and, during that time, has provoked considerable discussion within our industry. I guess the first question to ask is what was the idea behind it? I don’t think it’s fair to say its introduction was to replace .TAR and .PAX file formats which were traditionally used by tape archive /backup software, but rather that it offered a simpler and more open approach. Pre-LTFS, traditional archiving software was more involved to set up along with being a closed format, I believe this turned some away from archiving on tape.
The overall idea behind LTFS, a self-describing file system which defines the organization of data and metadata, you don’t have to depend on independent software to translate the contents of the tape into readable and usable formats. Consequently content and metadata can be accessed independently from the archiving software or database, as the standard file system view of the data is stored on the tape.
LTFS describes the files and folders on the tape in a familiar manner; this is displayed differently depending on your OS. The host machine, be it Windows, Mac or Linux, uses the tape’s file system to display the contents of the tape just as it would if you had a USB drive connected. Users can drag and drop from the host machine to and from the LTFS tape all without any additional software.
So what’s not to like about LTFS? It’s simple to use, self contained, multi-platform and the need for complex archiving software is negated.
In some applications though, the lack of file management can be problematic. Have you ever tried to manage multiple external hard drives used by a variety of users containing substantial amounts of different data? You typically end up with a wide variation to the file structure that you initially suggested to try and help with its management. It’s a fact of life that users prefer to use the top level of the drives and aren’t always willing to place files into designated folders. If they do there’s always a chance they will be abbreviated or spelled wrongly or, worse still, they’ll use characters that cause system administrators very long nights!
So while LTFS as a format is very appealing, it does lack the management that traditional archiving software provided, for example in spanning tapes, searching, replication and report generation.
I’m not sure the archive vendors greeted LTFS with open arms but, as the market has moved on, they have responded. We are now at the point where we have a good open format tape file system that independent archive vendors are integrating into their advanced archiving software. From my experience, all the major archiving software vendors have either already implemented, or are close to releasing, LTFS within their products and that represents a significant step in the right direction. LTFS is here and it’s going to stay. After all, LTO is a mature tape storage format with LTO-6 around the corner offering 2.6TB with a data-rate of 160MB/s.
This is not to decry disk archiving which is still favorable in some instances, but tape is cheap and compact and, when sitting on shelf in a library, is not drawing power. It will be interesting to see how Sony’s standalone Optical Disc Archive fairs against LTFS & LTO-6.