Storage

01:58 PM
Howard Marks
Howard Marks
Commentary
Connect Directly
Twitter
RSS
E-Mail
50%
50%

The Data Rich Are Different--They Use Tape

I recently had the pleasure of attending a press and analyst day with the good folks from Spectra Logic, where the state of the market for tape storage finally sunk in. While most corporate users have evolved their backup processes to reduce, or eliminate, their reliance on tape, users with lots of data still find tape to be the best solution. F. Scott Fitzgerald once said, "The rich are different," and that's true for those who are rich in data, as well.

I recently had the pleasure of attending a press and analyst day with the good folks from Spectra Logic, where the state of the market for tape storage finally sunk in. While most corporate users have evolved their backup processes to reduce, or eliminate, their reliance on tape, users with lots of data still find tape to be the best solution. F. Scott Fitzgerald once said, "The rich are different," and that's true for those who are rich in data, as well.

For decades, tape and backup were inextricably linked. Tape was removable, providing a simple way to transport data, and affordable in the days when any disk in the data center was a precious resource. As the cost of capacity-oriented hard disk drives fell and corporate IT types learned to trust them, disk-to-disk backup systems, especially with data deduplication, became more attractive, especially in the mid-market where managing tape drives' quirky need to ingest data at a large fraction of their rated speed and managing stacks of tape cartridges without detailed procedures were problematic. Today I usually recommend that organizations backup to some disk target and reserve tape for a more archival role.

Today, it’s not backups but folks with really large quantities of data that are driving the tape market, especially those with data that, unlike backups, doesn’t dedupe well, if at all. As a result, big tape systems are selling well into markets like video, including both surveillance and production; scientific and other high-performance computing; and Internet companies where a petabyte isn’t considered big data. After all, Google used tape to restore users' data after the last Gmail failure, and I’ve seen recent reports that Google is the world’s largest consumer of tape for data storage.

EMC’s biggest backup target, the DD890, only has 768 petabytes of disk space. Add in that a petabyte of idle 3-Tbyte disk drives alone will use 2 to 2.5KW of power, even before power supply losses, controller fans and the like. However, a petabyte of tapes will sit in your library without consuming power, so you can see the attraction of a big tape library.

The work flow in media and scientific data centers is quite different than in the corporate world. When a camera crew from "National Geographic" media (which sent a speaker to the event) comes in from Patagonia or Bora Bora, their footage is sent directly to the tape archive to make sure there is a safe copy. Only then is it retrieved and sent to edit stations on the floor.

SpectraLogic’s T-Finity redefines the term "large tape library." This brontosaurus of a system can hold 50,000 LTO-5 tapes, picking and placing them into 120 tape drives. This sucker is so big--almost 100 feet long with all 40 frames installed--that each of the dual robots has a webcam and a light that projects its position on the data center ceiling so techs can see where they are at any given time. When we went on the factory tour, they only had a 31-frame model set up because that was the biggest one they could fit in the room. They’re tearing down a wall and moving the employee gym to make enough space to build out a 40-frame model. If one T-Finity isn’t enough, the company can install a pair of skybridges to connect up to eight of these monsters together, passing Spectra’s 10 tape Terapacks back and forth between them. Luckily, nobody’s called saying they need more than 3.2 exabytes in a single library complex yet.

While you may be phasing tape out of your backup process, the big boys still count on it.

Howard Marks has no business relationship with Spectra Logic. Spectra Logic did pay for travel to the analyst event

Howard Marks is founder and chief scientist at Deepstorage LLC, a storage consultancy and independent test lab based in Santa Fe, N.M. and concentrating on storage and data center networking. In more than 25 years of consulting, Marks has designed and implemented storage ... View Full Bio
Comment  | 
Print  | 
More Insights
Slideshows
Cartoon
Audio Interviews
Archived Audio Interviews
Jeremy Schulman, founder of Schprockits, a network automation startup operating in stealth mode, joins us to explore whether networking professionals all need to learn programming in order to remain employed.
White Papers
Register for Network Computing Newsletters
Current Issue
Video
Twitter Feed