Predictions for 2006: The storage saga continues

It's that time of the year again: The festivities are closing in, the New Year is knocking at the door, and everyone committed to public writing feels compelled to dust off the crystal ball and impart their predictions for the coming year. This time, in addition to coaxing predictions out of my magic sphere, I'm going to make suggestions about how to cope with what's coming.

Storage security will remain a key topic in 2006, with vendors slowly adding more features to their products. As we all have read in the news, the lack of built-in security features has proven to be particularly damaging for data stored on tape media, especially when the tapes leave company premises or are lost (d'oh!).

Quantum is the first vendor to announce a comprehensive road map to strengthen security features for tape drives, libraries, and media. The new features will include hardware locks, electronic-key management, and support for SSL, native encryption, role-based access, and audit logs.

Quantum will initially focus on its latest value line tapes and on the upcoming successor to the SDLT600, but it's easy to predict that other vendors will soon be considering similar security measures for their tape devices and media.

As for suggestions for other storage segments in 2006, such as networked or host-attached arrays, I can't help feeling like a glutton in a pastry shop. There are so many interesting new products and technologies in the works that it's difficult to decide where to start.

For example, next year we could see 10 Gigabit Ethernet become more popular and push iSCSI performance higher. Clustered storage is still a very attractive segment and has some intriguing solutions in the pipeline, such as the modular Storage Brick by Terrascale. With its exabyte-size (yes, I said "exabyte" -- that's more than a million terabytes) file system based on SGI's XFS, the Storage Brick could gain popularity with companies needing top performance, virtually endless scalability, and easy management.

Also, SAS (serial attached SCSI) is finally expected to take off in 2006. I predict that SAS will take over most of the new storage shipments, but after a recent (sobering) conversation with Shaun Walsh, Dot Hill's senior director of marketing, and other storage mavens, I see the move from parallel SCSI happening in slow motion rather than at full trot. In fact, one of the latest updates from Dot Hill delivered Ultra-320 SCSI connectivity for its SANnet II box -- not quite SAS yet.

Unfortunately, SAS won't be the only promising technology moving at a slow tempo in 2006; companies both large and small have other priorities, keeping them off the cutting edge. Given a choice, for example, many storage admins, would rather choose an easy way to stay compliant with regulations than have a faster, larger, trendy gizmo.