Savannah Peterson: Hello and welcome everyone. Thank you so much for joining us and being a member of the Portainer community. We have an absolutely thrilling interview for you today and a conversation that I personally am super excited to have, because I think edge computing is extra cool. Today we are joined by Bryan Rodriguez who is the principal platform architect at Intel working on their edge team. Bryan thank you so much for being here with us today. How are you doing?
Bryan Rodriguez: Well, thank you for having me.
SP: This is such a treat. Since I started working with Portainer a year and a half ago one of the most exciting applications for it was edge computing, and so getting to be here with you today I feel a little bit of my nerd fan girl going off in my chest, which is very exciting. And when we were warming up in the green room - you've known the Portainer team for a while. Tell me a little bit about how the relationship developed.
BR: Yeah, so Intel is always looking for something to demonstrate the art of the possible with their hardware right? And a lot of customers don't really care about what's under the hood. And one of the things is as Portainer was that sort of ‘judge a book by cover, see how it really looks’, and when we started trying to decide what's out in the market Portainer was pretty much it. And it was pretty easy to use and get going, and it was easy way to actually display some of our cool features. We approached Neil (Neil Cresswell, CEO Portainer). We had some ideas, we did a statement of work, and then voila here we are with the edge type architecture.
SP: So, four years ago essentially. I mean really at the beginning of this program. How did you find Portainer initially?
BR: Well we were in the docker containers and it just seems a buzzword right? Throughout the forums there was “oh I’m using Portainer”. It's just me, you know the way I like how Neil says it humanizes your interaction with the command line. And of course, that's what it does. And Portainer just makes things a bit easier on the learning of what containers and docker is. So that's it. It was just a buzzword in the forums. And he said one day, “why don’t we connect? I just emailed back, and we met in San Francisco in a conference and ‘voila, here we are’.
SP: I love that. We're sending the cold email folks, just as a side note and pro tip from Bryan there. I love it. So just so we get everybody on the same page, if they're not quite as deep in this space as we both are, how do you define edge computing?
BR: Well, this is a very broad topic. You can Google this and get so many different answers and I’m gonna get a little geeky about it, okay? For me, really, it's about the cost of what it is to send data over the wire that separates this. So another analogy if we have LAN which we're just separated from over a wireless network and there's a cost associated to sending all those packets of data back, right? And it can be very expensive over 5g 4g 3g. Well imagine if there was zero cost and the bandwidth is unlimited? Then you wouldn't have a separation of LAN, WAN and edge. But that's really for me what defines the edge - is this cost prohibitiveness, this amount of data you can send over the wire. And that starts separating you from your cloud infrastructure, from your edge infrastructure, and as you start moving away to those small IOT devices that you want to collect analytical data and turn it into data? That's the edge. That's really what it comes down to - how can we be proficient and cost effective at capturing human interaction in the world and translating that up into the cloud, which is becomes aggregate data and analytics. And it really comes back to what we can sell you in an ad, or you know, how when you walk into a store, it does a better job saying, “oh you've been here before right? Those are that's what the edge is to me.
SP: So, let's dig in there just a little bit. What are your favorite edge use cases?
BR: Honestly, in more robotic and government plane type use cases, I find a little bit more enticing and intriguing. We had some vehicles too, because these are little islands of their own micro data centers if you will. And you've got a ton of them. And if you're in a vehicle you know there's a lot of stuff going on - reading your face, reading what's going on in the car for example. And it has to work autonomously, and it's separate from the cloud, and it has to function and still do a job with safety in mind. And by providing a real business need. So, to me, isolated micro data centers of 7 or 10 nodes inside vehicles or robots, or my kind of thing. Yes retail stores and venues are very cool and they have some real applications there. But for me I get the most joy in this robotic sort of self-intelligent autonomous type use cases.
SP: And talk about an instance in which real-time data and the ability to transmit fast and efficiently is literally life critical. I mean that's such an adventurous use, right? I appreciate that. So, you're venerable in this space - how long have you been working in edge?
BR: Well I would say the real first time was 2015/2016, and edge was kind of a word but we recognized that this world of capturing human interaction and turning that analog environment into something digital to translate it into dollars very early on. And we were trying at Intel to capitalize ‘how do we enable our ecosystem to be able to deliver to our actual customers a platform to yet give them the same cloud-like experience, but at the edge? And one of the biggest pieces the operating, the op-ex cost, the factor of installation deployment, we need to develop things that are more ‘take-it-for-granted’ type attitude, where it just works. And these things don't exist right now because we've had a lot of collateral, and unfortunately AWS, GCP, Azure have made this world so easy for us to develop software, and we have to apply it to the edge. And that's really the part that we have to strive and help our customers with - reduce the operating costs for scale.
SP: Why is that so relevant to do right now?
BR: Because (this is just my opinion about the world) I think culturally we're becoming more instant and stuck on instant gratification, and we really don't care about how or what engine’s under the hood. Or, “I just want my application to work” and “I just want to make some money off this as a consumer, but I also as a person being using IOT, I just don't want it to work.” So in order to be able to scale with thousands of IOT devices we have to remove the headache to just installing software and managing it. And one of those key ingredients for me is Portainer. Portainer eliminates a lot of headache cost associated and I can manage a lot of devices from a single point.
SP: I love it. We are all here for those types of endorsements. Now I agree with you. The connection of instant gratification I think is really relevant, and when you're used to having such an instant cloud experience and you want to apply or have a similar UX farther away, literally. Or however that plays out. It's so important. You're touching on why it's such a challenging time for the user right now and why your timing and our timing I think is so succinct, and why the teams are working together so well. Very well summed up. Well getting geeky with it, and going farther down that path, what are your thoughts on Kubernetes at the edge? Is it valid or is it docker still leading here? What's going on?
BR: Good questions. I’ll say this. Kubernetes design pattern is needed for the edge. How it's implemented could be a little bit adjusted and some of that's where I believe Portainer’s coming along. There are particular use cases with Kubernetes - the design pattern works. I said all over the place but in using Kubernetes with some heavyweight hardware like Xeons or something that was high competition but runs beautifully. As you start getting out to the remote far edge, Kubernetes overhead starts to cause some problems, and so that's where I look at Portainer trying to evolve Kubernetes, and Portainer trying to meet those demands. Again, honestly customers at the end of the day don't care that it's Kubernetes. They just see that ‘I need to be able to do my job, deploy my work, and get what I want done’. And if Kubernetes is an ingredient in that whole stack, fantastic! But the reality is that what customers are viewing, and system integrators are viewing is a control panel and then the end user the actual customer who's buying and seeing, is totally something different. They really don't see Kubernetes.
SP: Right. I think we're all trying to simplify the complex tool necessary to accept our technological advancement in this space. I mean Intel Inside is actually an example I use a lot for how things like this in the future will just have been built in. And you'll know it’s in.
BR: And like you said - Intel Inside. But the thing is you don't care that it's Intel at the end of the day, you can get your job done.
SP: Right? Just such a good marketing campaign. It was early in my formative marketing years. I get it, I get it. I use that all the time with the deep tech stuff and blockchain and whatnot in our world now. So what are some of the pain points for customers when they're managing edge deployments? Like when the device goes offline and you need to access it remotely, or what do you see?
BR: Well, this is the part where I think people do care from a system integrator - why it's Intel. But there is an extreme cost associated for having somebody physically go to a location and fix a problem and so one of the key words we talk about is out-of-band manageability. What that really entails is ‘I’m not using the software that booted up this machine to manage it. I’m using some other mechanism outside of there to restart remote manage fix it, recover from where we are’. So one of the advantages that Intel tries to employ in their hardware is this capability of, even if the operating system totally crashed I can still get connected through an out-of-band management capability to administer that. And one of the things that I love lately that Portainer has done, is that it has included some of these features that Intel provides to enable out-of-band management in the case of a situation of a failure. See there's a smart city use case where you've got devices on utility pole. Physically getting up there is a pain and is costly, so in this scenario out-of-band is a big deal, and it's very cost effective to be able to add a central location anywhere from the world, and be able to administer, figure out what's going on, restart it, or even create self-healing architecture to actually do that on behalf of a human, to try to recover the device.
SP: It's pretty amazing. It's almost like remote surgery.
BR: Yes it is. I’m excited about new technologies coming out to even rebuild the entire machine out on that utility pool without even a human being there.
SP: I think we'll get there. I’m very excited. There's reason to be optimistic about this kind of stuff. It allows a lot of scale and a lot of learning in a way that is cumbersome right now that I think will become much more fluid, and it is fantastic. So as a team, you've given us a lot of different examples. I can imagine it's challenging to prioritize. What are your priorities as an edge team?
BR: Well it's an excellent question. Fundamentally, priority wise we're just trying to create a product that is easily consumable, so that our customers can write their business applications to run on top of it. Really, we're doing our job when our customers take for granted the hardware and the software they're running on. Literally, if they come back and say ‘it just works, it's not a headache’, we're winning. So our priority is, how do we enable our system integrators and our partners to simply provide ultimately our bigger customers that ‘take-it-for-granted mentality’ on our hardware. So that's what we prioritize. How do we deliver, onboard, provision those little things, and get to the point of just being able to deploy your application, just like your mobile phone. You simply install an app and it just works. That's the intent.
SP: I love it. Intuitive design with the user in mind!? How human-centric is Intel?! (laughs). This is so nice and it's warming my heart to hear this because it shows why our collaboration is such a compatible and fluid one. Because the priorities both from a utility as well as from a mission perspective are incredibly aligned. so I’ve got to ask, what do you think about the two new features we've built in partnership together?
BR: Well I’m a little biased but I love them, even on a personal note. Even geeking out, the fact that even if I was not working for Intel, I can take the software and manage devices that I wanted to if I had home projects or something. Right out of the box, I’m able to onboard devices, remotely manage them. It's extremely powerful. And then on top of that, deploy software on there. These two new features add quite a bit of a new capability from a Portainer perspective and it really opens up the market from an Intel effort, to sell and get it in front of people these out-of-bound manageability capabilities for small IOT devices. So absolutely. It also reduces operator experience. This operating cost of simply, ‘how do I onboard a thousand devices? And simply, I would love it if I could just buy something off Amazon or ship it over UPS and it arrives, and the person installing it, pulling it out of the box, they simply just plug it in instead of having to do all these processes. People don't realize the amount of effort and cost of what it takes to set up this hardware in venues and in stadiums and vehicles and just let it run. It's expensive. And so the whole objective here with these features of Portainer and Intel is to remove that cost, so that customers can get to what they want to do to make money off their applications that they deploy.
SP: Yep, just trying to make the magic happen, and just trying to facilitate. Why is working with companies like Portainer versus building it yourself important?
BR: I would say that Intel is very good at making hardware that we take for granted, but we are a large company. We should also enable our ecosystem, our partners to also be very good at what they do, and together we can make something beautiful. And it has a similar outcome for our customers, so yes Intel could do all those sort of things, but let's have Intel focus on what they're really good at and let's have companies like Portainer focus on with really what they're good at, and we have these two ingredients and we make something beautiful together. So absolutely, elevating all, everybody rises together in these scenarios.
SP: I’m so with you. I’m a community builder, so you know. And it's so true. We move so much faster and gracefully together when we are able to serve at best scale and pull from the strengths of others, especially with something like Portainer. And working with a company at scale like Intel, it's a really cool synergy. I just love it. What do you wish that people knew about edge right now, or what do you think are maybe some myths you'd like to dispel, if any?
BR: Well, for a particular geeky audience, I will say this. What works in the cloud does not necessarily work at the edge. We've been burned so many times, trying to take technologies that work for the cloud applied to edge devices, and it just does not work, because unfortunately these edge devices could be - like I said - on the utility pole. One side could be 130 degrees and the other side could be 80 degrees and they could fail. We don't have this nice data center-like environment, physically and logistically infrastructurally. I made up some words (laughs)
SP: (laughs) We're building the future, we can build the lexicon.
BR: Yeah there you go. But from consumers wanting to get into this thing, I would say this new paradigm is going to happen regardless of if you're concerned or if there's conspiracies behind it. But the world of taking in an object detection and facial detection and vehicle detection and all this, is just gonna happen and it's actually gonna make your lives a bit easier to just do things and operate. And there is a certain thing you will benefit out of all this coming out of here, so enjoy it and have fun with it. And there are things to do.
SP: Yeah, it's not all bad. I think conversations like this I believe work to help make the future less scary. It's only scary if you don't understand. And folks like us know the existing limitations right now so there's really not a lot to be afraid of yet. I think that's awesome. Do you think that edge will become the prevalent deployment mode in the future?
BR: Back to the original statement about whether when you abolish the WAN versus LAN sort of scenario. Because of capitalism, I don't think there will ever be free internet, and edge is always going to be a thing, because you have the telco world wanting to make money in between, and you have the cloud folks. So I think though you're going to have a symbiotic relationship and heavily influenced. And you know, a lot more money is being put into edge and it's going to grow to where they live with each other. But different ways of revenue is going to be achieved out of it. So heavily together type architecture is moving forward, and it's going to shift. And if I were to make a prediction it's going to be at least 50/50, maybe a little bit more 60 in the edge market and 40 in the cloud. So that's where I would say it would go eventually.
SP: I like it. That's a strong prediction. And that actually brings us to our second to last question. What are your future predictions? Anything you'd like to share with us from your snow globe?
BR: Well I know this is very lofty but mankind to really get to go to the mars and explore beyond space, we need to build things that we take for granted. We need to feel like we can depend on autonomy and robotics and these things, and edge is an avenue to discover these capabilities in autonomy in devices. And as we start growing this, we could build trust in the things that we need to help us survive in the future. And I really believe this is a foundation and a motto we should take, where we build things with trust to be able to explore and expand mankind.
SP: I’m not even going to talk too close to that because I just want to let that stand for a second. I love that, that's great. It's so fun asking people what they see in the future because you get a taste for how broad their mind goes. And for some people it's a very succinct 18 month prediction, ‘we will shift to this technology’. And I love that you just went straight there. I’m a huge space fan. This interview has been out of this world and very valuable Bryan. I can't thank you enough for your time.
BR: Thanks so much.
COMMENTS