Data Tape: Dying a Slow Death or Already Dead?
July 18th, 2013 By: Industry Perspectives
Jeff Yaptengco is a Systems Architect at En Pointe Technologies, a national solution provider of IT products and services. Follow @EnPointTech on Twitter for news and updates across the technology industry.JEFF YAPTENGCO
En Pointe Technologies
Over the past 10 years, numerous articles have reported that “tape is dead.” Yet in 2013, many data centers are still using tape. I’ve been in the data storage industry for years, and our customers STILL have some sort of tape storage lingering around. Data tape is not dead; however, its purpose has changed.
In the past, linear storage media was used as the first line of defense to restore lost data. Yet today, there are few backup administrators using this technique. Today, many customers are using tape as a duplicate backup. They run quicker to disk within their backup windows and can be duplicated or archived to tape for off-site storage.
Additionally, they are using tapes for their second, third, or fourth line of defense. There have been many fault tolerant systems like replication, mirroring, snapshots, clustering, and backup to disk that have superseded tapes as the first line of defense against data loss.
Evolution of Storage and Backup Technology
From reel-to-reel to data tape, the technology has evolved over time. When I started, quarter-inch tapes (QICs) were used, which had about 200 to 400MBs of capacity. The savings incurred by backing up to tape over disk were astronomical. Best of all, tapes could be moved off-site in case of site disasters.
The technology grew to larger capacity media formats like DAT, AIT, DLT, and LTO and faster tape drives were created. Companies like Archive, Sytos, Maynard, Connor, Colorado, and Seagate popped up everywhere and were competing to manage tape and drive backups through schedulers and databases.
By this point the demand for tape storage had grown so much, tape automation robots from manufactures including Exabyte, Sony, Quantum, HP, Compaq, IBM, and ADIC were used to move hundreds of tapes in and out of tape drives. This allowed the automation of moving tapes to span beyond their capacities. They had a fixed price, and a fixed amount of slots, which could hold a fixed sized tape. When slots were empty it equaled wasted cost. Today, tape automation robots are being used less because of their limited expandability and cost over disk; only a handful of all those manufacturers remain.
Data Tape Killers
The first and most obvious data tape killer is that disk prices have decreased exponentially. Almost all data centers today use fault tolerant “Nearline” systems (also called High Availability or HA) like disk RAID and mirroring, server clustering, and snapshots. If these systems are working properly, restoration from backups may never be needed. For example, mirroring usually involves writing the data to two forms of storage at the same time. One is active while the second one is passive. If the active storage or disk fails, the second one can take over with no down time. With these systems, if disks are lost, the first line of defense to recover is not from tape anymore, but the “Nearline” resolutions.
Another “killer” was the creation of Virtual Tape Libraries (VTLs), an emulated tape library system that writes directly to disk. VTLs looked and acted just like a tape library, but were composed of disks. However, modern day backup software can now also backup directly to disk storage via local disk, OST, NFS, and CIFS protocols.
Backup software today can backup to disk storage with de-duplication, a compression technique used to get single instances of data on the storage, saving possibly 50-90% of capacity on repeated data segments. Additionally, with cloud services, it doesn’t need to be stored locally in the data center. Backups can go off-site to cloud storage providers, and with de-duplication, only unique data needs to be sent to the cloud.
The Future of Data Tape
So what is the future of data tape? Well, “How important is your data to you?”
Even as a second, third, or fourth line of defense, tapes are still the most cost effective way to move data off-site and store it away for long periods of time. Additionally, a few of the data centers still using tapes require data to be essentially cataloged and locked up in a vault never be touched.
There is also a matter of compliance requiring data be kept for a certain amount of time. There will always be data that must be kept or converted from old backups on tapes. For companies where data can, at no cost, be lost, tapes will always hold a purpose. They will always be portable and store large amounts at lower costs. So while it appears the purpose of data tapes has changed, tape still has a long life ahead.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.