Have you heard of Panzura? If not, you are probably not alone. The company is relatively new to the DevOps market. But it’s not some new startup that grabbed some VC money by using DevOps as a buzzword: Panzura has been around for years, perfecting its cloud-based file system that can deliver eye-popping results in terms of speeding file requests, as well as the volume of data that you have to transport. For DevOps, this kind of speed savings is just too good to pass up.
I had a chance to sit down with Rich Weber and Barry Phillips of Panzura and discuss why DevOps teams are finding Panzura’s offering so enticing. Barry is the VP of Marketing at Panzura, but don’t let that title fool you—Barry is highly technical. Rich is chief product officer at Panzura and is also really knowledgeable. Together, we had a great conversation about this interesting offering and why the DevOps market actually came to Panzura rather than Panzura coming to it.
As usual, the streaming media player is below, followed by the written transcript of our conversation. Have a listen or a read, or both. Enjoy!
Alan Shimel: Hi. Alan Shimel, Editor-in-Chief, DevOps.com here for another DevOps Chat and I’m happy to be bringing—I, well, I think introducing to our audience a new company or a new company to the DevOps face perhaps that we haven’t had a chance to speak to before and that company is Panzura, and I’m happy to be joined on this episode by Rich Weber, Chief Product Officer at Panzura, and Barry Phillips, Chief Marketing Officer at Panzura. Gentlemen, welcome.
Weber: Nice to be here. Thank you, Alan.
Phillips: Yeah. Thanks, Alan, great to be here.
Shimel: Okay and just for our audience, though, Rich, that was you first and, Barry, that was you second. So guys, I’m gonna guess right now that our audience may not be familiar with Panzura, can one of you give me the elevator pitch or a quick background on Panzura and we’ll go from there?
Phillips: Yeah, it’s Barry. I’ll do that. So yeah, so Panzura and it’s interesting as I start this off, most people may go I’m not really sure why you’re in this DevOps space, so what we do is have a hybrid-cloud storage solution and really what that means is we put all of our customer’s unstructured data in the cloud and we happen to cache the data that they’re using on physical or virtual appliances that’s set in the offices. And that happens to work really, really well for some specific use cases because of some unique technology that we have inside of our software so you know one of those is for manufacturing and it really makes CAD applications work really, really well across sites, architectural apps, engineering apps and just like those use cases we’ve have customers actually bring us into the use case into the DevOps use case and really specifically being able to globally distribute software builds and artifacts quite quickly. And so that’s really the part of DevOps that we play and it really kinda plays into the continuous development, continuous delivery, continuous integration space and essentially what happens is in the case of electronic arts you know we dramatically taking the time it takes to distribute globally you know 50GB build files from 10 hours to essentially minutes and we’ve got many different examples of customers like that, so it’s always great to be brought into a specific space by a customer need that’s taken a platform that you have and really using it for a specific use case.
Shimel: Absolutely. And, Barry, of course when we’re talking about a time difference of 10 hours to a few minutes, I mean that’s the kind of eye-popping differentials that we like to tell in the best of cases in DevOps, right, where we’re putting those kinds of savings, those kinds of numbers in front of people. So thanks for that background. Rich, if I can ask you so you know obviously there’s use cases here for the DevOps audience, for the DevOps market and DevOps use patterns, as we say. As chief product officer, Rich, what kind of—you know, how did you steer feature set? How do you plan your road map to kinda further help the DevOps community with this product/service?
Weber: Well, I mean typically it’s based on feedback from real customers that we have doing workflows and things that run through there and like so that eye-popping one that you were talking about is how do we reduce that time by so much? You know a lot of that came from working with software developers and one in particular is in the gaming industry where imagine their builds are very large, right? They’re building into like Blu-ray ISO images and things and enabled low-cost QA centers to test that. Now you can have anywhere from 20GB to 50GB resulting builds and transferring those takes a long time, right? Suddenly, you come to the point of, ‘Hey, I could probably get that data across the globe faster on FedEx than I can digitally over the line.’ Our product really helps do that by providing a global file system that’s available in each site and by virtue of talking to the Panzura device it’s automatically in the cloud. It’s automatically duplicated, compressed and transported and available in their real time. Especially if I’m doing many hundreds of builds very quickly, the deltas between those builds are very short so effectively I can make those available for tests, production and immediate use through my build process.
Shimel: Excellent. And, Barry, you and I spoke a little bit offline and you know I think I had asked you about so is this a product or service that is really only for big companies or for unicorns you know of the world? I mean who’s a good candidate, especially from a DevOps point of view, for Panzura?
Phillips: Yeah and I would even extend it, I would say, beyond DevOps, right? So if you think of any company that does any sort of global testing and development whether that’s pure DevOps, whether that’s somebody who is in the process of putting DevOps in place or even somebody who’s basically not even there yet that when you think of just the ability to move these builds and artifacts quickly that really increases the number of test cycles you can do on the software itself. Of course, that’s better quality software you know typically gets market—it gets software to market faster. It’s all those typical things. But I think it really encompasses anybody who is actually has dev and test teams essentially around the world.
Shimel: Yep. And another kind of phenomenon that we’re—that I’ve observed, Barry, in talking to a lot of people, obviously, is this idea of a multi-cloud. So let me back up. When we first saw the whole sort of cloud revolution, if you will, I think the—everyone agreed that hybrid cloud was the way to go. No one was moving a hundred percent of their infrastructure to a AWS or Azure, a public cloud platform, but there was a hybrid where some of it might be public. Some of it might be in a private cloud in a data center. Some of it may be on-prem, right, but it was that hybrid situation. Now we’re seeing multi-cloud, right? And I think it’s a different animal. Barry, Rich, what do you think?
Weber: I think multi-cloud is good and particularly for DR purposes but going to that hybrid concept that you were speaking of before and how do companies really bridge the, hey I’m working inside of my four walls or inside of my data center and how I take advantage of that cloud space that’s where Panzura really helps companies bridge that gap. So in particular, if I am working within my own data centers and my own sites and I’ve got multiple sites, that example that I gave you before not only was the time reduced from many hours of days down to minutes but in terms of that footprint that went down to one—from 1.5PB to 45TB by utilizing the cloud because I had one central location of the data and I didn’t have to store and move and transfer many replicas. So that cloud gives you the ability to aggregate and make that data available quicker. Multi-cloud I think gives you an opportunity to then further do a disaster recovery even redundant across clouds not only from sites but from cloud to cloud, right? That first step is really getting that data and aggregating it in the cloud itself and Panzura really gives companies the advantage on how to move there and take advantage of that and even burst inside of computing, not—and beyond data, as well.
Shimel: Absolutely. I mean and that’s exactly where I see with multi-cloud. Once I’ve been able to move that data to the cloud whether I have some of my infrastructure on Windows, let’s say in Azure and some other stuff on NOS and AWS and maybe something on OpenStack at Rackspace or somewhere doesn’t make a difference ‘cause my data is rather—you know my data is multihomed anyway or not multihome. That’s not the word I’m looking for. Panzura takes care of the data. I don’t have to worry about that. Is that—is that the gist of it?
Weber: Yeah, I mean, in fact, we’ve got customers who I would say maybe you didn’t start out as being cloud-forward customers or cloud-savvy but Panzura makes that simple, right? If all they have to do is make a mount, my workflow stays exactly the same as it is today and you don’t have to worry about where that replication it goes, what data center is getting used. Panzura makes that part simple. All you got to do is mount it regardless of where I’m at. It’s the beauty of having a global main space in a global file system.
Phillips: And maybe to further add on top of that, Alan, is you know Rich mentioned workflow and as you think of you know companies who essentially go from you know waterfall to Agile and then get into you know really this DevOps model. You know we can sit underneath there and it stays the same for them ‘cause we are just we look like a regular storage mount for them, so everything will remain the same as they move down this process into DevOps.
Shimel: Great. So let me get a little geeky on us. Rich, can you kinda describe what the Panzura stack, if you will, looks like? I know you referenced a Panzura device earlier. What exactly—give us a little—our audience a little bit more detail on that.
Weber: Yeah, so I’ll try and keep it simple but on the front end it looks just like a file server you know whether that be like a net app or even a Windows file server on the front end, so it integrates with AD or other authentication methods and you just mount it, right? So, on the front end you’ve got either an SMB/CIFS front end or an NFS front end. You mount that drive and you store your data there. As data comes to us, it runs through our stack and our stack begins with de-duplicating and that de-duplication what you need to know about Panzura is it’s not just for the data at that place. It’s for the entire footprint for all data written everywhere because that de-duplication information is part of the metadata that’s shared globally. Then it gets—after it gets de-duplicated it gets compressed, then it gets encrypted and none of the customer keys or security access stuff exist in the cloud, so none of our customers actually even have to trust the cloud provider themselves because no data sits in there in clear text. Then it all gets compressed, packaged up and put in the cloud and that’s done at the block level from a de-duplication standpoint. So as it run through our stack we integrate WAN optimization, as well, so moving that data becomes very fast and simple and on the other end everything works in reverse and the customer doesn’t have to manage all that; all that happens automatically. Again, all they see is a drive mount and access the data in exactly the same way they would from their filer today.
Shimel: Has anyone ever used the term data storage as a service or data storage and access as a service?
Weber: Yeah, absolutely –
Shimel: Would that be fair?
Weber: Yeah, I think you especially as you move into that cloud paradigm, right? You’ve got platform as a service, storage as a service, software as a service and this is where, like I said, Panzura makes the adaptability of that service very simple, right? That’s storage and object storage in the cloud without you having to write specific apps to do that. Panzura allows you to adapt those type of either storage as a service or platform as a service integrated into your existing data center or workflow.
Shimel: Okay and guys, if I can ask ya, can you give our audience a sense of how this is priced or sold or offered or however you want to call it?
Phillips: Yeah. We essentially sell controllers and those controllers are what sits in somebody’s office and that can be either a virtual machine or it can be a physical controller and the more capacity those controllers have on them the more expensive they are. The other thing we do is we do have controllers that run inside the cloud and so and they would run inside Amazon right now and we do have just a preview that’s running inside Azure and the whole idea there is now not only does every single office see the same file system across every office but also that cloud does. So if you happen to have some sort of application or service running in the cloud, let’s say it’s platform as a service and you’re looking to be able to access file data that may be on premise, you know right now what we provide there is the same exact cloud data everywhere. So in all those cases somebody would purchase a controller from Panzura and that would be either a physical controller on site, a virtual machine that they would put on top of their own hardware, or it would be a Panzura controller running inside the cloud itself.
Weber: Yeah, I mean it depends on what your needs are. I mean in our hardware system think of this as a lightning-fast flash filer that sits inside of here, where our algorithms concentrate on the cache or very fast local performance usage. But software-only deliverable inside of a VM for flexibility of the customers or, like Barry said, an in-cloud solution if I want to burst into the compute and have my data there, Panzura controller in there, we’ll help ‘em do that. And again, global file system and mail work in concert together.
Shimel: Got it. Guys, I’m curious. As part of maybe your market research customer feedback, what have you, what has the feedback been from people from DevOps teams, let’s say? I don’t want to use the term DevOps engineers. It gets people up in arms but you know from teams that are using DevOps within some of your customers have you gotten any feedback from them on this and is it different than your customers at large?
Weber: I think the interesting feedback that comes through is probably twofold, right? There’s two things that you hear. Most of the concentration really comes in productivity, right? It’s about that speed. How do I make things faster? How do I get bills done quicker and get product to market faster? And we give some examples of that and I’m not kidding. I’ve been into customers before and they say, you know, “Hey, can you beat FedEx bandwidth?” and it just blows me away. You’d think it was the nineties or something again. So when you’re dealing with big files that helps and it’s not only about delivery of the final executable but also about doing builds in remote locations. If I’ve got remote mounts on those build artifacts or dependencies, people hit builds for private builds and remote site and wait ‘til the next day. With us they hit a build and they enjoy the same performance that they do at the headquarters or where the files reside. The second portion is, of course, the economic side, right? How do I save money? And I think that most of the feedback comes in as they pick us up or productivity gain, get products back to market faster and then they found out without having to do replicas and manage multiple file loads of data they end up saving a ton of money and it’s another example I gave you, right? 1.5PB of replicas down to 45TB—that’s a huge storage savings.
Phillips: Yeah and bear in mind I’m the marketing guy here, I mean the things that I’ve heard is when you think of these distributed teams who are working in a DevOps environment it really makes them seem like they’re really in one office, right, because it makes that speed seem just as fast over a wide area network across the world as essentially on a local area network in one office.
Shimel: Sure. And, Barry, you’re being facetious when you say you’re just the marketing guy. I happen to know you have two degrees in computer science so let’s not hide behind the marketing title too much. Anyway, gentlemen, we’re—actually, we’re past our time already. We always go past our time and I apologize, but I need to ask one last question which is the question we usually end our DevOps chats with. Though there’s two of ya, I’m gonna ask you just to recommend one book to our audience that they really should read as soon as they can or at least put it on their list and get to it when they can. What book would that be?
Phillips: Yeah, it’s Barry. I really enjoyed reading, “The Big Switch,” from Nicholas Carr and it really was—it was interesting ‘cause it—and I think it applies to DevOps and especially with DevOps in the cloud. I think someone wrote it’s kind of the whole chocolate and peanut butter you know together as a chocolate peanut butter cup and just really gave examples of what happened when power used to be generated from waterwheels and how that moved to the grid and really making that comparison of how things are moving to the internet. So I thought it was a great book and I recommend it for anybody.
Shimel: Great. Rich, you second that one?
Weber: Absolutely. [Laughs]
Shimel: Great. Well Rich Weber, Barry Phillips, Panzura—of Panzura, thank you for joining us on this episode of DevOps Chat. We welcome you to the DevOps.com community and we hope to see and hear more from you in the near future. Continued success and we’ll see you both soon.
Weber: Thank you.
Phillips: Thanks, Alan.
Shimel: All right. This is Alan Shimel for DevOps.com and DevOps Chats. Thank you and we’ll see you again at our next DevOps Chat.