Feeds and speeds
LTO-5 provides 1.5 TB of uncompressed capacity at 140 MBps native throughput. This represents a capacity increase of about 2x but a speed increase of only 15% over LTO-4. When I asked IBM's Bruce Master, a member of the LTO Consortium, about this relatively small performance increase, he said it was intentional. First of all, the 2x capacity increase has been maintained between generations, but speed increases had slowed to 50%. LTO-3 was 80 MBps, and LTO-4 was 120 MBps. Master said that after LTO-4, the group "did a lot of research with end-user focus groups, analysts and manufacturers and determined that for most applications, users weren't saturating their LTO-4 drives. And while they did want more capacity on each tape, increasing speeds too much would just add unnecessary cost to the drive." In the future, the consortium expects users' throughput needs to increase and it plans subsequent LTO generations to resume these 50%-per-generation speed increases.
The LTO roadmap currently extends out through LTO-8. It lists LTO-6 at 3.2 TB of capacity and 210 MBps throughput, LTO-7 at 6.4 TB and 315 MBps and LTO-8 at 12.8 TB and 472 MBps; all specs uncompressed. According to Master, the current "disk-first-then-tape" backup strategy is more likely to drive performance requirements in the future than the traditional direct-to-tape model did with earlier generations. Although disk backup certainly isn't new, LTO-5 will see more and more use cases like writing snapshot copies to tape, as well as image backups common in VM environments, allowing users to catch up with current LTO performance levels. Applications for large-file reference data, like rich-media archiving, are also likely to push performance requirements.
LTO-5 supports media partitioning, in which the drive can write two variable-length partitions on each tape, which can be used by backup and other application developers to add features. The first major implementation of this is the LTFS, which IBM developed and has made available as a free download. LTFS runs on Mac OS and Linux (Windows support is expected in Q3 of this year) and uses the first partition to write an index and the second for data. This can greatly improve access times and overall data management but also make the tape a "self-describing" tape file system, meaning it can be read by any computer that runs the LTFS software. In a backup application, this means backups could be written to an LTFS tape and read anywhere, without requiring the backup software.
There are other interesting implementations of this file system as well. "One could imagine a 3 TB thumb drive," said Master, essentially a 3 TB file system, that's cross-platform and portable. Another option could be a "tape NAS," or an inexpensive file server that could be set up for reference data that doesn't need the immediacy of spinning disk.
Write once, read many (WORM) functionality and tape encryption were added to LTO-3 and LTO-4, respectively, and both are included in LTO-5. They aren't currently supported in the LTFS but are on the roadmap.
It's clear that archiving is a primary use case for LTO-5. In fact, LTO-5 and LTFS were announced at the National Association of Broadcasters trade show last month, to the delight of attendees. If you look at the data situation that the broadcast and imaging-based industries have to deal with, you can understand the reason for this positive reception. High-definition content and broadcasting, 3-D movies, and large overall increases in image resolution have exploded data sets by an order of magnitude or more. With significant cost advantages (with prices as low as 5% of the cost of D5 broadcast-format media)and its ability to support tape automation, LTO tape is making inroads into these industries and seems to be even better positioned to continue in that role. In fact, several hardware and software manufacturers have launched a consortium called Active Archive Alliance, to encourage the development of tape-based archive solutions for these industries and other more mainstream IT data archiving applications. LTFS would seem to support that objective nicely.
What LTO-5 means for VARs
- The end of "tape is dead": It would certainly seem that tape's not going anywhere. While dedupe and off-site replication have reduced the role of tape as the primary target for backups, enterprises of all sizes continue to offload data to tape at some point in their data protection process. In fact, according to a survey conducted by the LTO Consortium in late 2008, 60% of managers operating disk-only systems said they intended to add tape back into their environments. This is probably due to a number of factors, such as power costs and availability, data center space, and the sheer economics of tape storage (especially with 3 TB LTO-5). Now, with LTFS and technology like file virtualization, a "tape tier" offers an easy way to store reference data and one more reason to look at tape. VARs have historically been moving away from backup in general and tape backup specifically, as a product category to lead with when they talk with prospects. While this may continue, to some extent, LTFS can give VARs something new to show customers.
- Archive is alive: ILM and the nearline concept had a tough time in the mainstream IT market. The software needed to manage and move the data was complex and there generally weren't many sources. Also, the value proposition was based largely on the high price of disk, but disk storage's continual cost reduction made for a relatively unstable environment for the new products that were needed. Thanks to the explosion of data in the rich media and imaging industries, as well as the power, space and cost issues of even low-end disk arrays, tape is again positioned as the best technology for large archives. Now, with LTFS providing a simple solution to store reference data inexpensively, tape archiving may finally be ready for the more general storage market. VARs should keep on top of the developments of software vendors as they integrate this new technology into their offerings. They may also want to explore the possibility of selling archive solutions and get in touch with the Active Archive Alliance.
About the author
Eric Slack, a senior analyst for Storage Switzerland, has more than 20 years
of experience in high-technology industries holding technical management and marketing/sales
positions in the computer storage, instrumentation, digital imaging and test equipment fields. He's
spent the past 15 years in the data storage field, with storage hardware manufacturers and as a
national storage integrator, designing and implementing open systems storage solutions for
companies in the Western United States. Find Storage Switzerland's disclosure statement
This was first published in May 2010