 Live from the Sands Convention Center, Las Vegas, Nevada, extracting the signal from the noise. It's theCUBE, covering HP Discover 2015, brought to you by HP. And now your hosts, Dave Vellante and Jeff Frick. Simon Watkins is here as the worldwide product marketing manager for HP Storage, joined by Tim Pite. Gentlemen, welcome to theCUBE. Thank you. Thank you. I was talking about one of my favorite subjects in the world is London. Yeah. It's a great city of London where you live. Born and bred. Born and bred. Awesome. So soccer fan, I presume? We call it football, but yeah, it's soccer. Yeah. And of course, HP Bristol, renowned for taping, we're going to talk about taping a little prop here. Absolutely. Hold that up, actually. Yeah, there you go. How much capacity? 6.25 terabytes on a single cartridge, LTO6. That's compressed, but a lot of data. So let me ask you a question about that, 6.25 terabytes. How long does it take to scan a 6.25 terabyte cartridge? So the transfer rate for an LTO6 drive is about 1.4 terabytes an hour. So you can work it out. That's the transfer rate of an LTO6 drive. Okay. So do the math on that. And then as that capacity doubles, which it will do, the time it takes to scan that tape won't change, will it? No. No. It's kind of forgotten. Let me ask you a question. What's the maximum capacity of a disk drive today? Was it 8 terabytes? 6 terabytes? Yeah, around that. How long does it take to rebuild an 8 terabyte disk drive? Quite a while. How long is it going to take to rebuild a 30 terabyte disk drive? Quite a while. A month? Two months? Three months? No. The point we're getting to here, folks, is tape is alive and well. Everybody says tape is dead. It's not dead. Absolutely. For reasons that I've implied in my little soliloquy there. But so anyway, we'll talk about some of that. So Tim, tell us about Barrett Jackson. We saw a little side of your car there. Hey, yeah. It's Barrett Jackson. We were a car auction. In 1971 in Arizona. And we have grown over the past, coming up on our 45th anniversary in January. And we have grown to become a much larger than just a car auction. We're an event. We're a lifestyle. Media sensation. Yeah, media. We're live television coverage. Which really is one of the main reasons why I had to look at the tape solution was for the television side of things. I have 18 years of historical footage. All HD, where do I put that? Having it sitting on online storage or spinning disc, it just didn't make a lot of sense. So we looked at the tape solution. So add some more color to that. So specifically, what are you doing with that? What are you storing on that tape? So we have our broadcast that we broadcast. We've been broadcasting for the last 18 years, like I was saying, on Fox. Now we're on the Discovery and Velocity channel. And that hundreds and hundreds and hundreds of hours that we have of footage going back historically, we digitized it. It used to be on tapes. We had it on many, many, many, many formats, you know. Beta cam, three quarter and DVC pro, all that. So we went through and digitized all that. And then it was a matter of where do we, where can we centrally put all of this information and we can access it as we need it. And we needed a large storage array of some sort. We looked at spinning discs. We looked at building a home system, a home brew system, just a bunch of discs, things like that. And then it came down to reliability. We wanted it to last. I mean, this is 18 years of footage. And we went through all this trouble to digitize it. Where we put it that we know it's going to be safe. And looking even further is having more than one copy. And the more reason I say looking forward is because now that I've digitized it to 18 years, now it's time to go back and meditate it and start going and slicing, dicing it up and making clips and making it searchable. So that, let's say we want to do a special on 65 Mustangs or 69 Camaros or pre-war classics, whatever it is, that we can make edit decisions very quickly by knowing that we have that footage. If we search it and realize we don't have it, let's do a different segment. So just being able to retrieve it if we need it and when we need it was really the big thing we're looking at. So this is your digital transformation, right? Essentially it's a metadata problem, metadata management challenge. Okay, and then I can see a lot of ways in which you could use that footage. Do you make that available to your community? And that's really, I think, the next phase. We need to delay the foundation first. It's been that process because we have an auction every four times a year. So what we do when we do an auction is we pack everything up, we move our data center to the event, set it all up. So when we have these opportunities for these projects, it's hit it hard, take a break, hit it hard, take a break. So now we're in that process of, okay, we've finally digitized everything. Now the next phase is cataloging it. So Simon, we've seen the repositioning of tape as the primary backup medium to now one of long-term retention. It's a superior long-term retention. We've written about this on Wikibon for years. It's the superb long-term retention because, as you were pointing out, it lasts longer than spinning media. It's far, far less expensive. The issue has always been performance, the perceived performance, but we can even put forth a case. We'll talk about that where tape is actually higher performance. But so give us the update, if you would, Simon, on tape, from HP's perspective, the technology, the market, the strategy. Yeah, so tape is obviously, it's a very mature technology. It's been around more than 60 years now. But we think that just because the technology is mature, it doesn't mean there aren't any opportunities and it certainly doesn't mean that there isn't any innovation. So in terms of the opportunity, the market was worth over $2 billion last year worldwide for tape. And also, secondly, more data is being stored on tape than ever before. So we're seeing more capacity shipped on tape. It's a double-digit growth, we're seeing, in terms of capacity shipping on tape. I think the reason for that, the reason why adoption of tape is remaining very strong, is as you said, tapes kind of blend of low-cost, high-capacity, removable, reliable storage, is allowing it to evolve from its traditional role as a backup medium to, as you said, more of a kind of long-term retention archive use case. Now, when it comes to what we call cold archive, tapes low-cost, less than 1 cent a gigabyte, the energy benefits of tape, it's a green technology, you store it off-site, it doesn't consume any energy. The reliability benefits of tape are very important. The drawbacks of tape when it comes to archive was it was traditionally not a good solution for active archives because getting to data on tape, you'd need to use an ISV application to write that data, and it was often written proprietary formats, so 10 years down the track, if you want to get that tape data back, you'd need the ISV application. That all changed with the introduction of LTFS, right? And that's an open-format specification with a store of content on tape. It makes tape self-describing, just like a USB disk, so that accessibility problem goes away, and also because you're storing content in an open format, you're not tied into any ISV application. LTFS is a real game chamber for active archives, and one of the really powerful implementations of LTFS is what we call Tapers NAS, and what that does is it virtualizes a tape library behind a disk cache, it could be flash, it could be spinning disk, and what that means is that applications can now access tape directly via that cache, and if you use flash, as opposed to spinning disk, you get the performance benefits of flash with the long-term cost and reliability benefits of tape. So that's kind of the value proposition, and that's exactly what Barry Jackson is using. Yeah, so there's been a big marketing push in the industry from a number of vendors that say, tape is dead, tape, S-U-S-C-K-S, blah, blah, blah. But the reality is for most practitioners out there is it's got a role, whether it's the backup of last resort, you know, the stuff in an iron mountain somewhere, or in specific use cases like Tim is seeing. So I want to explore this a little bit. I went pretty fast at the beginning, but people watching are probably thinking, how can Dave say that tape could actually be higher performance? So let's talk about that a little bit. So there are only, I don't know how many disk drive vendors are left, it's a handful, two, three, four. But nobody's investing in, I mean, high performance disk is an oxymoron. You can't spin it any faster, and you can't put more track, more data on the track because the heads aren't evolving. They're hermetically sealed, nobody's investing in heads anymore. And so as a result, getting data out of a disk is like sucking data out of a thin straw. You don't have that problem with tape, right? Because LTFS gives you advantages, but the tape, the heads in theory, I guess you could stagger them. And so that's why I said earlier, the time it takes to scan or whatever it is, a six terabyte cartridge today, is going to be the same today as it is 10 years from now. Five years from now, whatever, when that capacity doubles or triples, the effective capacity, the bandwidth rate of a disk drive is going to go into hell in a handbasket. So now it starts to open up all these new opportunities to use tape in new different ways. So I want to explore this with you, Tim. So you've got all this archive footage and you're starting to metatag it now, right? Where are you at in that process? So what we're at in this process is it's really kind of a, it's almost like a three-step approach. The first was digitized, we've got it digitized. We now are ingesting it into a media asset management system. What that simply does is just kind of start creating the database, file names, size, things like that. The next phase is going to be... Excuse me, that's a piece of ISV software. Is that right or is that something you guys wrote? No, no, it's a, if you don't want me sharing it, it's a company called Squarebox and CatDB is the name of it. I want to know because we have the same problem here. Cube has a lot of video. Yeah, and I do our live stream too, so I'm going to be using this system for that as well. And so what we're doing now that we've ingested it and I'm doing it kind of in phases. I don't want to jump in and throw 90 terabytes at it because that's basically what I have to ingest right now. I'm kind of doing it slowly and learning and really trying to know everything about it. So now I'm in the process of creating a proxy version. I've got 1080i footage. Some of those files are 400, 500 gig. Well, with CatDB, what it allows me to do is allows me to make a browser-based searching functionality. So if I have an editor in New York that needs a footage for let's say a charity vehicle that sold the last auction, I can have them log in, search the file, and then send me edit choices, which then sends me an XML file. I then put that into our edit system and it will then retrieve the high res version. Whether it's sitting on tape, whether it's sitting on the disk cache that we were talking about earlier. And that's an automated process? It is an automated process in the sense of once you have the metatagging done and you export a very simple XML file. I bring that into Final Cut. CatDB talks with Final Cut and knows that well, here's where the footage is. The Q-Star manages which tape it's on to Q-Star, it sees everything that I throw at it. So whether it's on tape only or whether it's also already in the cache. So Q-Star indexes everything and then goes and gets it when you need it. Right, so let's say I have something that my cache is 10 terabytes, by the way. So let's say I have something that was 50 terabytes ago and I go to search it, it pulls it up, it shows that the file exists. It's going to take, you know, 15, 16 minutes to pull that off the tape and load it into the cache. But what that allows us to do is still make edit choices. At the end of the night you run all your XML by the time you come in, it's sitting on disk cache. So hopefully by the time overnight it was able to go and pull all those files that you need on the high resolution side. So the metadata about what files live where historically it would sit on the tape, cartridge itself, correct? As far as the file management portion is concerned, is it for the tape library? And it still sits there, in your case. And then we also have another layer of metadata which is the video content itself. So could you elevate that file level metadata to a cache, a flash, let's say? That's technology that's in development. That's the flight you guys have written about, right? David Floyer, my colleague, who's also from England, coined that term. But the concept being elevate that metadata, file metadata to a flash layer. And then write algorithms that, because it's all, time to first byte is going to be faster on disk. But time to last byte is going to be faster on tape. If in fact you can write algorithms and say, okay, I need this data and I'm going to pass the fifth secret request, I'll grab that on the way, reorder it at the back end. They're not like rocket science algorithms. Smart people can write them. And that blows away the performance of disk for large object file retrieval. Like video. Like video. Or even shorter files that you're storing for long-term attention that you can concatenate. So we think the tape has this wonderful future because it's way, way, way less expensive. And it's, believe me, in this scenario, it's higher performance. So it's a real winner. So that two billion, we think, is going to be flat to up even potentially. And because also there's a lot of runway for the aerial density for tape. So we think that cost per gigabyte is going to go down, right? So the cost advantage of tape, I think will only get better. So disk in this scenario kind of gets squeezed, the spinning disk. It's not the high performance. It may not be the best bit bucket long-term as technology vendors work on these. So is HB actually actively working on that type of technology? Yeah, that's good because there's only a few tape vendors out who have the capability to do it. You're one of them. And we're one of them. But we're the market leader actually in the mid-range space. We're still investing a lot in tape. We've got L7 coming out soon. So we've got a good roadmap moving forward. So yeah, it's a good business product. So what's your reaction to this conversation, Tim? Presumably you'd like to see the industry sort of move in this direction. I do. In the sense of, so I've talked about this 90 terabytes that we've digitized. That's a scratch of the surface of what I have. I've only digitized. I was going to say 90 terabytes must be nothing. Oh, no. Because all I've done was digitize the broadcast. We still have every single camera. I mean, I see here you have a four-camera setup. Yeah, right. If I so choose chosen, I could then do premium archive footage that's never been shown on television. So the possibilities are endless. With the tape system that we went with, it's modular. So I can keep growing this thing. I don't know you tell me indefinitely. Yeah. I mean, that's the thing is by the time I outgrow what I have and my need is, you know, he's already made the decision. So how do you automate, you automate the categorization of all this data, right? And that's what Q-Star or CatDV does? It's more the CatDV side. So I was explaining to Simon earlier about that, was I know everything about the vehicle that was selling, you know, from the year that make the model, the engine size. I know all that information. That's on our internal system that runs the auction company. We are looking at a couple of different options there, both moving forward for future recordings and historical recordings. Right now, XML, it seems to be the most viable option for historical. But then we're also looking at return on investment. So in other words, let's see what the need is and how much we actually really need to go and meditate every single vehicle. Do we need to make this publicly accessible or so? Then absolutely we do. But again, going back to where we're more than just a car option or an entertainment brand. So we have some segments where the talent on air talent will be talking about a vehicle and speaking with the owner and the historical family significance of it. Is that interesting to put out there in the world and make it searchable? Or is it all about the cars? Well, it's just a really interesting use case. I mean, this is a great example. If you've got content, that content is an asset. Can you monetize that? Maybe, maybe not. Can you create value for your community? Can you launch new businesses? I mean, I'm sure these are the discussions you're having internally. Like every company should be having. And so, but that's a, you know, you kind of fall into it as technology evolves. You've got this corpus of data. What do you do with that? And it's not an easy thing to attack, but the technology's there to do it. And the ideas are floating, which I'm sure they are. HP's enabling those ideas we heard and the idea economy. So that's fantastic. Well, we're out of time. We've got to leave it there. But I'll give you the last word, Tim. Sort of advice for practitioners, peers trying to sort of struggle with this challenge. What would you do differently if you had to do it over again? Yeah, no, specifically to my case. Really, I guess the only thing that I would do differently, which I luckily have enabled myself to go about it differently, is a larger cache. You know, because I'm dealing with such very large files, more cache is definitely better. You know, if you're using it for more of a user level, you know, document smaller files, you know, you don't really need as much of a big cache. But when you're dealing with 500 gig files, one after another, you chew that cache up very quickly. You know, and that's really, I think, just to your research. And I think if you really look at the cost per gig and growth, you will start really seriously, because I was anti. I'll be honest. Yeah, no doubt. A lot of naysayers. This isn't going to work. And eventually performance. Mark my words. Cache is king where you heard it here, you know, for the millionth time. But Tim and Simon, thanks very much for coming on theCUBE. Great segment. Appreciate it. All right, keep it right there, everybody. We'll be back with our next guest right after this. This is theCUBE, we're live from HP Discover. We're right back.