The newest technology in tape, LTO5, is making news lately for both its performance and its impressive capacity per piece of media. IBM recently announced a Long Term File System specifically designed for LTO5. The idea is that the file system enables tape to become a better medium for data archiving and not just backup.

Of course, adoption of LTFS is going to be driven by both ISVs and the makers of the operating systems. After all, if the intention is to use tape as a “drag and drop” drive, then an O/S is going to have to mount that tape and present it as a file system just as it would any old disk.

But what folks like us are more interested in is the new partitioning feature. Since tape is linear (Linear Tape Open), it must be written sequentially, and ideally, at the maximum rate the drive and tape supports. This makes it ideal for fast streaming backup targets, but makes search and discovery, a key component of archiving, a bit difficult.

We agree that tape is a great archive medium, even given the challenges of the linear format. Seven10 has been archiving file system data to tape for 5 years, albeit with a proprietary file system and through our tiered storage product, StorFirst EAS. Even older technologies without partitions can be used with intelligent tiering and caching.

Today, our customers can tier SAN disk and tape for a mirrored WORM archive file system. Or, they can simply use a locally attached disk to cache data written by users and set policies for how large the cache should be before flushing it out to tape.

Our CIFS/NFS presentation and highly scalable virtual file system takes care of the bits and bytes of media support for them. Want to mirror that tape and bring to another site for use? That’s absolutely possible by simply inventorying the mirror at another site and then reading the data from the share. Our read caching even enhances 1 + n access read times.

The advent of the partitioning on tape has been thought of internally for years now. There is even preliminary support in the product today; we were just waiting for the appropriate physical technology to work with. We’re really excited about the possibility of making search and discovery even quicker than it is now.

Repeat after me: Tape is not dead.

- Adam Marcionek, Principal Engineer for Seven10