Future Markets with the Potential to Thrive on LTFS

Making the most efficient use of the media and technology

LTFS 3 Future: Segment 1


This is Jay Livens and we're continuing our podcast series talking about LTFS and how the technology was created, some of the key elements of the technology, and now we'll talk a bit more today about the future of the technology; where we think it might be going.

Now joining me in this podcast is my guest Michael Richmond who is an expert in the area of LTFS and was part of the team, in fact one of the leaders, in creating the initial specification for LTFS, and the initial designs back when he worked at the team at IBM.

So Michael, thanks for joining us again. Thanks for coming back to the podcast again here today.

Hi Jay, it's great to meet you virtually again to continue this conversation.

Great. Well let's jump right into it then. As we've talked in the past of a number of different questions that we had covered and I want to dig in to a few of those in the context of LTFS and where it's going in the future. Now in the last podcast I asked about what data types might be best for LTFS and I want to expand upon that little bit here. It's not just about data type but do you think there's specific markets that might be best served, either today or in the future, by LTFS technology?

There are certainly markets that are potentially more suitable to the benefits that LTFS provides. These markets are primarily about what the size and type of data that is commonly used in the particular market. So for example, with all of the LTFS development work and the initial go to market that we did at IBM, we are focused on targeting the media and entertainment market, and specifically broadcasters and the videographer's market. This is largely because in the broadcast space, they have very large files that they need to store, and they need to store them for a long period of time because the video files are their primary asset in the M&E market. That characteristic of needing to store large volumes of data in relatively large transactions for a long period of time really speaks to the forces that LTFS brings to the table in terms of a data storage format. And there are a number of other markets that have similar requirements. For example, there's the oil and gas industry generally collects very large amounts of data when they are surveying a potential exploration site. We've seen some interest from that industry for using LTFS because this industry generally does a large scale survey, pulls together the data, analyzes it, and determines whether it is worthwhile to invest exploration in that space but then needs to store that data because as technology evolves, an analysis that may determine that a potential field is not worth exploring, that decision may change in 10 years’ time so being able to go back and retrieve the data to do a more modern analysis is often a requirement in the industry.

On a similar scale but in a more fine grain market, video surveillance is potentially very interesting on LTFS. You could potentially construct a video recording system that has a single LTFS drive, you drop the cartridge in once a week or once a month, depending on what your data rate is, and your video cameras just feed and lay down to the single LTFS tape. You can then choose to put that tape on a shelf or you can overwrite it. All three of these examples I'm talking about storing large amounts of data that needs to be stored for a long period of time. This inevitably raises the question of potentially the backup recovery space that we've talked about on an earlier podcast and there's also another area that is potentially of interest which is using data tape as a near-line storage for small work groups. One of the benefits of LTFS and using data tape is that you can build a relatively large amount of storage in a small volume if you're willing to trade off instantaneous access time and shift your access time to be on the order of maybe a one minute access time.

What I'm talking about here is the potential of using a 2U tape library or a 1U tape library which might hold 20 cartridges in the 2U library and the tape drive and stuff that full of LTFS volumes which will give you about on the order of 35 terabytes of storage in a 2U rack space using LTO-5 tapes at a very cost competitive price compared to using an array system. With LTO-6 and future versions of the LTO specification, the amount of storage you can fit into such an enclosure will continue to increase.

That's a broad brush strokes look at potential markets. All of it's really talking about needing to store fairly large volumes of data that's unlikely to change over time so you don't want to be using LTFS for small transaction based systems because LTFS, although it will work to record small files, it's not the most efficient use of the media and of the technology.

The Speakers:

Michael Richmond
Jay Livens