Elevate the Edge
Elevate the Edge
33. Edge to Cloud Integrations with Keith Townsend
Elevate the Edge welcomes guest Keith Townsend, Founder of CTO Advisor, who discusses Edge to Cloud Integrations.
Guest Bio
Keith Townsend Principal at the CTO Advisor a research and advisory services firm. Keith brings his experience covering 4-decades in IT focused on technologies like infrastructure, storage, virtualization, cloud and edge to the podcast.
Hello and welcome back to the podcast elevate the edge. I'm Maribel Lopez and I am joined here with my co host Jo Peterson. Hey, Jo.
Jo Peterson:Hey Maribel.
Maribel Lopez:And we're super excited to be here today with a friend of ours who is also an amazing technology leader. We're here talking with Keith Townsend, he is the founder and principal at the CTO advisor, a research and advisory services firm. Keith brings multiple decades of it focused experience in technologies like infrastructure, storage, virtualization, cloud, and of course, what we're talking about today the edge, we're going to be talking a little bit about edge to cloud and integrate. missions. And Keith, welcome to the program.
Keith Townsend:Thanks for having me. It's been a long time in the making, right?
Maribel Lopez:Absolutely. But you know, we've all been busy traveling the globe going to different things. I think the last time we tried to talk to you, you were in Barcelona. With VMware, we've we've been at reinvent together since then, AWS is reinvent. So once again, thrilled to have some time with you.
Jo Peterson:Great to be here. Hey, Kay. So we're I'm going to get started here. As an IT infrastructure subject manager, a subject matter expert, I should say, what are you most excited about when it comes to the edge?
Keith Townsend:The applications? You know, it's the technology is all about applications. And infrastructure is built for applications. As we're building out the infrastructure for the edge. I'm looking at apps, I'll get into a advisory conversation with someone who is leading audio visual company and the use cases are starting to pop up that I don't think about him. He was asking me about containers at the edge for AV, and I'm thinking AV equipment and containers. I don't know, let's let's have the you know, break down the use case. And let's talk about it. And it's it has some really interesting things as we give users and end users, they're the capabilities, how they're taking it and taking it in directions, I would never have thought of.
Jo Peterson:Well, having been part of the AV Club in high school. It was the only containers we passed around or ice cream containers. This a little different than what you're talking about. But I hear what you're saying lots of excitement there. I know Maribel is got a burning question. She's ready to ask you.
Maribel Lopez:Yeah, so. So Keith, you know, there's been a lot of transition in the past few years, and you've been helping a lot of businesses migrate from traditional IT infrastructure to the next generation of computing, which was around public and even in some sense, private clouds. And I'm wondering what kind of advantages you when you speak to businesses you see them having as they're looking at moving to the next step, and towards adopting the edge. I suppose we could even say what do you consider the edge, but maybe you could just paint a picture of some of the advantages you see businesses moving towards with edge computing.
Keith Townsend:So let's say you have an ESG, you're a farmer, you have an ESG goal. And then this is ESG is something that we're all getting super interested in to going into 2023. And you're able to take woodchips and burn wood chip in a bio friendly way for energy. That just sounds like not possible. And I talked to a customer who's doing exactly that 200 acre greenhouse, and they have IoT devices on IoT IoT sensors on each individual plant. So let's take tomatoes, and they're able with the sensors to determine if a plant is too big. If it's getting the right nutrients, et cetera, get into a long conversation about kind of how they're efficiently growing these plants in Canada, the IoT devices are kicking off 25 gigs of data per plant per growing season. So you're looking at about 37 terabytes of data every three to four months. And I asked the question, well, how are you powering these IoT devices? Because that seems counterintuitive. You're you're you're using electricity and energy to get IoT sensors and all that like you know, we're completely carbon zero because they take the woodchips to the leftover to burn energy and then refeed that back into the ecosystem as co2 for the plant food for the plants. So when you're thinking about edge, these are the type of applications that I would have never imagined. For the edge, and we're seeing edge and agriculture and edge and AV edge, and retail and food to table or farm to table is just amazing what we're seeing.
Maribel Lopez:Yeah, I'm with you on that we're seeing some great use cases with, you know, the cold supply chain making sure that they yogurt arrives at the right temperature so that it doesn't have any strangeness to it. We've seen pharmaceutical companies put centers in, you know, really getting COVID vaccine and make sure that it was at the right temperature. And those are other examples of edge. But I love edge and farming, because it's one of the things that people don't really naturally associate the edge with, right? Because it's very tech, and they typically think something data center, you're not necessarily something that moves, not something like rose. So there's all kinds of interesting aspects we're looking at for new world order with the edge. Jo, back to you.
Jo Peterson:Well, you know, you got me thinking here, Keith. And I want to see if we can tease through this a minute. I want to think about one of the OG examples of edge, which was Chick fil A, right? Use containers at the edge. And the whole idea was to find out when clients were coming into the queue into the drive thru, and then make fries, their wonderful, wonderful fries, or good fries, right? To meet the demand. But if we kind of extrapolate that a little into an edge cloud scenario, and think about maybe that, that data flowing back to HQ, about how much fries they need, can you sort of paint a picture for the audience about how edge to cloud might work in that scenario, and why you need both?
Keith Townsend:Yeah, so let's think about the what we're actively trying to do at the edge, which is provide the hot fries. You know, after this hour, go to DoorDash place, or for some french fries, and a Italian beef sandwich, which is a fabulously good sandwich here in Chicago, if you've never had an Italian beef sandwich in Chicago, checkout, Portillo's and grab an Italian beat have it dipped will hot peppers, you won't you won't regret it. But if you think about the logistics of that order, from a H perspective, this is not even meat going into the store, this is the store needing to receive the order and time the delivery of those fries so that when I get them, they are still relatively hot, that is a challenging computer science problem. So not just Chick Fil li at the edge or in the store. So you have the store, you have the demand in the store, you have the demand from on the cloud. And that's like a non traditional way of thinking about that problem. So data coming locally. And then if you go back even to the Chick fil A problem, even you know, take out the cloud ordering part of it, you want to do data analysis at the edge, you want to know you want to tell the fry guy or gal I was a fry guy at one point went how much french fries to drop when to drop them. And then you want to aggregate that data so that when you open up for delivery internationally or nationally, you can project what you know from a you can feed the algorithm more data. And you want to do that munging at the cloud where where you've aggregated all of that data, and then sent it the sandals updated models back down to the store so they can get even better at predicting and controlling kind of their, their, their workflow.
Jo Peterson:You painted that beautifully. Thank you and you went right to where I wanted to go, which was managing the data and the infrastructure. Talk to us about the lack of a plan. And while that might just upset the whole apple cart for lack of walkers can keep with food terms here and the risks that you would run by not capturing that data correctly and driving business insight.
Unknown:So let's even go back to the original use case that we saw the the farming use case 25 gigs of data coming off a to offer tomato plant. You want to be able to control how much water it gets to that tomato plant. I will say the tomato panel with gases watered once or twice a day with a certain amount of water is getting given a certain amount of nutrients 25 gigs of data multiply that by a few 100 plants. How do you get that data back if you're collecting all of that data, how do you get all of the data back up to the cloud, you can't, like that's a lot of data. So you need to be able to do analysis at the endpoint. If you don't, if you fail at it, then you won't achieve the business goal, which is to provide the nutrients at the time that the plant needs to nutrients in the water at the time that the plants the sunlight that the plant needs, at the time that it needs, and it will be too late and you won't achieve your goal. Goals. The opposite is true as well, if you're not getting enough data up to the cloud, you can't get these new world data scientists this company nature fresh, give them a shout out, excuse me, nature fresh, give them a shout out, they're doing a fabulous job of taking growers, people who grow up in agriculture, em turned them into data scientists so that they can now use their domain expertise at data science to that to get this expertise. But if they're not if the if the data isn't at the place where they can manage big data in the cloud, the these talented folks can't make that transition.
Maribel Lopez:Keith, I love where you were going with this in the sense of one of the things we always talk to organizations about is you're not doing an edge computing strategy just for edge computing sake. Right? You're doing it to support a business outcome. And so you actually teed up nicely with the farming example, some of the business outcomes and how much data you would need versus if you have no data. If you have more data, how to think about that. As you talk to organizations, I recognize that each industry is different. But are there a few maybe common pieces of advice or guidance or pitfalls to avoid? That you would like to share with the audience?
Unknown:Wow, this is a that's a big question Maribel because they don't
Maribel Lopez:ask big questions here.
Keith Townsend:So there's a bunch like right sizing the edge. And you would think this isn't that difficult, but it is really, really hard. How much I have a limited amount of power envelope that I can deploy at the edge, I only have so much CPU, so much memory so much, literally DC power? How do I keep these endpoints powered? When I you know, gave the example of the of the nature fresh farms? One of my big question was how do you powering all of this data analysis? At the edge? How are you getting the data from the edge back into the cloud or the data center for centralized processing? So there's these basic infrastructure questions? Do I put out ARM processors at the edge? Then that next level of infrastructure related? Problem is? How do I do command and control? So when then I know you folks have talked about you lazy? I've talked about this on the podcast before? How do I update the firmware at the edge? Like that's a practical problem, when I have a new algorithm that I want to put out, pushed out to the edge? How do I do the command and control for that? Am I going to adopt Kubernetes? That's what Chick fil A did? Or are these devices too small? Are these small ARM devices and I can't run a full container stack and I need to figure out how I'm going to do the development? And then that's the next layer of the problems like how am I going to do the development? Do I want a past type cloud like experience? Or am I going to do low level embedded system application development? These are some really difficult questions and there's not a one size fits all for IT industries. Some industries will you know automate automotive, they there's a whole set of standards that are different from you know, am I getting the data off the transmission or off the GPS system to completely different OSs and, and data networks, etc. So these are really big problems. But in general, you want to consider what's the size? What's the size and fit for my devices at the edge? Am I going to go with commodity off the shelf hardware? Am I going to spec something that I'm going to have built? Do I have the capacity from the chip manufacturer to deliver what I need consistently over the next few years? The lifecycle of that stuff? How am I going to services services at the edge and how am I going to get that data back and how am I going to control those? Those devices is out to the edge. Those are my infrastructure specific views. Once you get into the application and integration with your SAP ERP systems, all that there's there's just so so much, you know, I think someone should start a podcast and talk about all of these problems.
Maribel Lopez:Absolutely, I think we thought of that idea. Yes. But what I really liked with where you're going is, you mentioned something that's really important. I mean, it was all important, but there's no one size fits all. And sometimes, you know, I think as analysts, the three of us on the call are frequently asked by the organizations we deal with to give us the playbook, the easy button, the whatever. And, you know, what you brought up really gets to know there's a lot of specific questions, you have to ask about your industry and about your specific business and answer those and then the right things fall out of that, but you can't just pick up some wonderful edge computing playbook and make that happen.
Keith Townsend:Yeah, we haven't even talked about health care, and the the regulations associated with with privacy information and non stop environments. Have you ever tried to build a non stop, eh device that is a crazy power envelope in just all other kinds of challenges?
Maribel Lopez:The other one that's getting me right now is Private 5g. Learning Networks is not an easy thing. No, in general, right. It's not easy for a telecom provider to run a network that the fact that you think you're gonna roll out a private 5g for your edge computing environment is beyond difficult to execute on unless you have a specific skill set. So that really means that you need to be very partner led on that one. So that's another thing that I've been thinking a lot about for the edge. Jo, back to you.
Jo Peterson:And my head spin in. This is such good stuff. Keith, we have to have you back on the podcast. You're such a delight. But we're going to end the podcast today by asking you a fun fact that is non tech related. And I wish the audience could see your amazing, amazing, ugly Christmas sweater, because that is a gift in itself. Please tell us your fun fact.
Keith Townsend:The fun fact that this not as surprising anymore, but I love to share it anyway. Yeah, it's hard to imagine, you know, Maribel gave this description of me having decades of IT experience so you automatically assume I'm an older guy, which I have a little bit of gray, but I have a 14 year old granddaughter. What let that soak in. Like when you see the either you know I have I've gotten the magic to anti aging or there's a really interesting story behind that. So I really encourage you as you see me and public asked me how in the world do I have a 14 year old grandchild?
Maribel Lopez:Yeah, folks, he's he's not looking like he has a 14 year old grandchild that you can't actually see Keith or have not seen him on social media. Let me just say that that looks a little impossible here. So here we are. And I love that that's a nice personal fun fact. So this is a podcast where we like to get to know the edge but we like to get to know the people behind the edge. Keith, thank you so much for your time and attention. And Jo is always loving, enjoying sharing the podcast with you. And this is us wishing everybody a Happy New Year.
Jo Peterson:Happy New Year guys.
Keith Townsend:Happy New Year.
