by Molly Rector
Addressing astronomical data growth and increasing accessibility demands
Over the last 10 years, “Data Growth Continues its Breakneck Pace”1 could be the name of a regular column in any storage publication. Massive data growth has become the norm, which means that data center strategies have to address constantly increasing storage requirements. Many technologies have emerged to meet this need, and most attention has been showered on disk technologies such as virtualization, deduplication, and replication.
One of the more recent advancements involves a proven veteran of the data center: tape. Industry analyst Jon Toigo writes that “one of the most exciting developments in storage last year [was] Tape NAS [network attached storage].”2 Advances in tape have moved it from primarily backup to the first and second lines of data defense. The key advancement: making tape look like disk.
With this, administrators now have the option of using a single interface to access all storage. With tape and disk displayed using a directory structure interface, users can retrieve data independent of the media platform. With the much, much lower cost and power usage required of tape, this is good news for IT professionals and chief financial officers.
An added advantage is that all data can optionally be kept online, by combining disk with tape as NAS. With roughly 60-second latency, users can access all data stored on tape. Instead of having to ask administrators to access older data, a time-consuming process, users control data requests with only minimal latency.
The how-to on using tape as NAS is simple: Applications and appliances are already on the market, and it’s simple to add Linear Tape File System (LTFS), which presents data stored on LTO and TS drives using a file system format.
LTFS: The Why is the How
The move to use tape as NAS has gained momentum from the 2010 announcement of LTFS. With LTFS, the drive writes an index describing data on tape to a known location. Prior to this advancement, tape’s linear format didn’t support this sort of indexing.3
On reflection, it’s clear that disk has the same problem: Data is written randomly on the disk, so the data must be tracked so that it can be reassembled. Disk reserves space in a known location for an index. Now, LTO and IBM TS1140 drives do the same thing using partitions.
Since indexing is essentially pointers, the index requires little space. And the index is on the tape, so the tape is self-indexing.
The LTO consortium adopted LTFS as part of the LTO standard. LTFS is open source, which means tapes written with LTFS use “a standard format so that data could be interchanged between writers and readers regardless of operating system, tape, tape drive or software vendor.”4 That, combined with the users’ freedom to access data stored on tape without requiring administrator intervention or a proprietary application, means that LTFS is ready to use out of the gate.
The LTFS format is just one approach to presenting data on tape as though the data were stored on a disk or flash. Sites with large data sets, such as those in the high performance computing and the media and entertainment industries, have been accessing data this way for years.
Implementing tape as NAS is
straightforward. Add NAS tape using:
LTFS applications Software Appliances integrating disk and an application Cloud-based tape as NAS
Tape as NAS: Using LTFS
IBM researchers created LTFS using open source code, which stores the tape’s “table of contents” schema in XML. The design of LTFS came out of work for the media and entertainment industry, and it is designed for a much broader set of uses.5 LTFS works across multiple operating systems, and extended versions expand on the original code to work with tape libraries.
as NAS: Using a Third-Party Application
With tape a central component of storage, and with the wider acknowledgement of tape’s many advantages, including cost and reliability, more companies are offering applications that provide a file system front end to tape, with or without LTFS.
Media and entertainment companies have used tape as a primary storage medium for years. The need for rapid access to broadcast data makes perfect sense, as does the use of tape as a repository for huge files common in the industry, with disk serving up data as it’s rapidly transferred from tape.
Companies such as FileTek provide applications that create a file system front end for tape, optimize performance to take advantage of tape, and work seamlessly with disk. FileTek also understands that technology doesn’t stand still, and includes extensive support for migrating data from tape and from disk. Since disk migrations occur much more frequently than tape migrations, handling both provides a significant advantage.
Tape as NAS: Using an Appliance
Appliances, such as those provided by Crossroads and Cache-A,6 offer another path, with an appliance (hardware and software) presenting tape as NAS. Both companies take advantage of LTFS. Cache-A also works with TAR, an open-source tape archive format that has been around for decades.
Above: An Implementation of Tape as NAS
Tape as NAS: Cloud Platform
Tape, although sometimes overlooked as part of cloud storage, is central to cloud’s success. Some cloud providers are willing to discuss the use of tape in the cloud. For example, Thought Equity Motion7 provides a platform used by media and entertainment customers, and about 95 percent of their cloud data is on tape. The rest is disk, which the company uses solely as cache. This platform takes advantage of the open source LTFS. Rather than worrying about proprietary broadcast applications, the data is format-neutral through LTFS.
New strategies and technologies typically filter into use over long periods, but it’s hard to see any other option, aside from tape as NAS, that makes more sense for storing large quantities of data. Tape as NAS provides more terabytes of storage per kWh, greater longevity, and greater reliability. Tape is also fast. As industry analyst Jon Toigo points out, “Certain types of files, such as broadcast or surveillance video, genome datasets, and others are actually delivered more rapidly from Tape NAS than from rotating media.” Tape as NAS, through LTFS and other tools that display data on tape in a directory structure, provides the ease of use and management that has been missing in the implementation of tape over the last decades — and now that problem is gone.
Molly Rector is the EVP of Product Management and Worldwide Marketing at Spectra
Logic (Boulder, CO). www.spectralogic.com
1 SearchStorage.com, "Data Continues its Breakneck Pace." Storage Market Research Reports. Accessed March 2012.
2 Toigo, Jon. "Storage Trends for 2012," White Paper, Data Management Institute, 2012. http://www.drunkendata.com/wp-content/uploads/2012/03/Hot-Storage-Trends-2012.pdf
3Richter, Rainier. "LTFS, E-mancipation for Tape: Freedom and Savings for File-based Workflows." White Paper, Media Technology Market Partners, July 2011. Available online from http://mediatechmarketpartners.com
5IBM Linear Tape File System with Thought Equity Motion's Cloud-based Platform." Webcast March 28, 2012. https://events.unisfair.com/index.jsp?seid=30435&eid=556
6 LTO Consortium, "LTFS: Setting your Videos and Data Free Now and For the Future," Webinar, March 15, 2012, http://event.on24.com/event/40/13/37/rt/1/documents/slidepdf/lto_ltfs_presentation_nab_2012_final.pdf
7IBM Linear Tape File System with Thought Equity Motion's Cloud-based Platform." Webcast March 28, 2012. https://events.unisfair.com/index.jsp?seid=30435&eid=556