It is funny how this industry changes almost from week to week. Sometimes there is so much activity going on you do not even know where to start. Other times (like during holiday shortened weeks such as this one), it is difficult to find anything really noteworthy to write about. In light of the fact that this week was a bit quiet from a news perspective, I wanted to reflect on some innovation occurring in the area of LTO-5 tape and how this might lead to a renewed interest in tape media in the years to come.
Part of the reason that LTO-5 tape caught my attention this past week was due to an email that I received as part of a listserv to which I subscribe. This listserv primarily consists of individuals who own records management businesses and are responsible for the back-and-forth transportation of tape as well as the short and long term storage of tape media.
The specific article being circuluated in this email recently appeared on SearchStorageChannel.com regarding LTO-5. This author of the email who found this article summarizes the points made in the article as follows:
First, it seems that to develop a long-term view about where this industry is going you have to differentiate between data that is stored for disaster recovery purposes on one hand, and data that is stored for archival purposes on the other hand.
Some enterprises may not yet actually make the distinction between backup and archival data, relying on their backup tapes to serve both the disaster recovery and data archival functions. But it seems that there is a growing distinction between these two types of data and that over time most companies will view them differently and have different sets of data for disaster recovery versus data archive.
For disaster recovery data, the first-line disaster recovery defense for most enterprises will be a combination of a) in-house disc-to-disc backup or b) e-vaulting (which could be viewed as simply off-site disc-to-disc backup).
These mechanisms can offer faster backup and recovery than tape – and are very useful for recovering from “minor disasters” such as accidental deletion of an important file. E-vaulting has the added advantage of putting the data offsite, protecting against a single disaster destroying all of the data. However, since both disc-to-disc backup and e-vaulting are “online technologies”, both are vulnerable to software corruption, viruses, etc.
For archival data – which data by nature doesn’t need to be retrieved frequently or in a hurry – it seems that offline and offsite storage will continue to be a natural choice so long as it remains a lower cost and/or more secure alternative than keeping the same data online.
Of course, the concern of the group on this email thread is whether or not most enterprises will continue to want secondary, offline backup in the form of disc-to-disc-to-tape to protect against these risks for the foreseeable future. Further, they are also trying to figure out what What the main drivers are for growth in archival storage – whether it is primarily litigation holds, long-term storage of data from research & development projects, or some other demands.
In the near term it appears the role of tape in enterprises is going to decrease. This conclusion is supported by a recent report from the Santa Clara Group that showed a 25% decline in tape drive and media revenues in 2009.
However LTO media was a bright spot in this report and this decline in tape usage may be only temporary until some of the new features found in LTO-5 start to work their way into the mainstream. This will likely take 2 – 3 years at a minimum since enterprises first have to deploy LTO-5 tape drives and cartridges plus software that can take advantage of one particular new feature in LTO-5 will also need to become more widely implemented before one might see an uptick in the adoption of tape..
The specific feature to which I refer is LTO-5’s new ability to support a tape file system. This could have a substantial impact on the adoption and use of tape for purposes beyond just archive and backup. While the main features around LTO-5 that have been promoted are its higher capacities and faster speeds than previous generations of LTO, LTO-5 also includes two data partitions on the tape cartridge so that it can support a tape file system.
These two partitions allows files to be written directly to a tape and read independently by an OS or application. One tape file system known as the Long Tape File System (LTFS) is already available at no cost from IBM and was specifically developed to take advantage of this new functionality available on LTO-5.
Some of the features of LTFS are admittedly pretty cool. According to IBM, the IBM Long Term File System, is the first file system that works in conjunction with LTO Generation 5 tape. This storage software reduces complexity in data management and access time through the enablement of a self describing tape that includes a simple file index. This helps to decrease tape, file management and archive costs while improving response time for new business needs. LTFS’s features include the ability for files and directories to appear on a desktop in a directory listing and users can drag-and-drop files to/from tape.
However it is unclear from IBM’s website whether LTFS will work on all LTO-5 drives or just IBM’s LTO-5 tape drives since IBM just lists the Long Term File System as available for TS2250 and TS2350 tape drives. That does not seem to preclude it from being used on LTO-5 tape drives from other tape drive manufacturers but neither does it explicitly mention support for them.
It also appears that the author of this article was present at the show where the introduction of LTO-5 and LTFS was announced as they were met with a warm reception by those in attendance at the National Association of Broadcasters trade show. (For those unfamiliar with this trade show, it is becoming a popular show for many storage manufacturers to attend, especially the ones that provide scalable, high performance clustered file systems and tape systems.)
While that may seem like an odd pairing, those two technologies address the specific needs that the broadcast and imaging-based industries have. High performance, scalable clustered file systems provide them with economical storage platforms that they need when high-definition content and 3-D movies are in production. Conversely tape systems provide them with the economical storage repository they need when these movies, shows and images are archived.
Video surveillance is also a burgeoning market that has been switching to low cost SATA disk systems connected to cameras. While video surveillance was once almost the exclusive domain of tape, disk in recent years has almost completely supplanted that. However now that LTO-5 supports LTFS, it will be interesting to see if tape can take back some of the ground it has lost in the video surveillance market.
As I conclude this blog entry, I have to be honest, if someone had told me at the beginning of the week that I thought tape had a bright future ahead of it, I would have questioned their sanity. But after learning more about LTO-5’s LTFS feature and how it can make it possible for users or applications to easily store data to disk or tape from their desktop, it casts an entirely new light on the future of tape and it will be interesting to watch how this LTFS technology plays out in the years to come.