Originating Author: Nathan Thompson
With an enormous amount of bad press surrounding tape (primarily submitted by disk-only vendors) it is interesting that customers continue to deploy it as a fail-safe “backstop.” I will expand on this point.
In its most severe form "loss of data" is equivalent to “loss of enterprise.” Given the dependency of organizations on IT and IT applications, the cost of downtime has grown to an average of $3M per hour (for Fortune 1000 companies). But what happens in the event of permanent data loss? It’s catastrophic - even potentially deadly when one considers the full impact on our information driven military, health care, energy and public services sectors.
Disk based backup has a significant role in IT. Labor can be saved, remote backups can be consolidated and user initiated fast restores all produce economic benefit. However, in rare events, disks systems are subject to hardware failure, application failure, hash failures, undetected bit errors, operating system failure, fire, flood, accidental deletion and even sabotage. Organizations will always depend on tape for a final backstop, guarding against the unthinkable—simply because the unthinkable is un-survivable.
From Spectra Logic’s perspective, we see tape as a growth market. There has been much market consolidation and many vendors have failed or are on the path to failure. Given our financial stability, and long term (non public, non-VC funded) orientation we have been able to invest and expand our market share for the last 29 years. Our customers continue to use tape. In all of our known cases, after an extensive process of technology investigation, none of our customers have migrated entirely away from tape storage.
From my perspective as CEO of a profitable and growing tape library company, I’m going to take this time to predict a few relevant trends in this industry:
- First, LTO has become the dominant tape format. Over the last five years we have seen the demise or near demise of Mammoth, QIC, DLT, AIT, SAIT, Travan, 9940 and SDLT. I predict that we will view the near demise of T10000, 9840 and possibly DAT over the next three years. IBM’s 3592 (AKA TS11x0, which incidentally shares its underlying technology with LTO giving it the cost and quality benefits of high volume production) and LTO will carry on through 2020.
- Second, there will only be three or four relevant library manufacturers in three years. The volume and price leader will be BDT, who will continue to sell its products through OEMs such as HP, Dell and Sun. The brand leader will be IBM, who will continue to take share and ultimately dominate mainframe attached tape. The technology leader is and will continue to be Spectra Logic - further adding to its intellectual property portfolio of library encryption, density, disk integration, Media Lifecycle Management, connectivity, reliability and ease of use. There may be a few companies servicing their installed base, but generally the supplier base will have shrunk to just a few leaders.
- Third, in three years the industry pundits will continue to predict the “death of tape”, as they have for the last 46 years.
Action Item: Continue to use tape as a backstop against catastrophic data loss. But be careful in your vendor and media selection. Make sure that the the standards you choose will be supported for years to come so that in the unlikely case you need to get to your data, you can.
Footnotes:
