37 Petabytes of IoT Data -Nature Fresh Farms VMWare Explore 2022
Transcript
(logo chimming) >> All right, welcome to, No, not, I was about to say Valencia, Spain. 'Cause I was just in Valencia for QCon back in May. But we're in Barcelona, Spain. Speaking, wrong place, wrong time. There's two Fira convention centers here in Barcelona. The other one is in a beautiful part of the city that we plan on visiting. I have with me Keith Bradley, VP Information Technologies for Nature Fresh Farms. Keith, welcome to the show. >> Oh, thanks for having me.
Like I said, I've seen you in San Francisco and we were there and I really, we got a great conversation started and we wanted to hook up, but our times didn't match. And I'm happy that we got to meet here in this great city. And I also want to tell you, you're not the only one that went to the wrong convention center the first day. (laughs) >> Oh, really? I feel better, but it was a beautiful spot. >> Yeah. >> Did you get a chance to go back and check out the space?
>> Yeah, we were walking around, we were touring 'cause we did the bus tour and I'm like, "Ah, here's the convention center right here. " And I'm looking like there's no signages, there's no nothing. >> Yeah, there's no VM, usually there's, you know, there's the big pop and- but there was no VMWare, VMWare Explore signs. " We start looking it up. I'm like, "Oh, wait a minute. " (laughing) >> So, we've had plenty of stories like that. A lot of language barrier but Barcelona's a beautiful city.
And I'm enjoying my time here. Speaking about beautiful things, Nature Fresh Farms. You stop by the booth at VMWare Explore in San Francisco, as you said, we got to talking about what you folks were doing from a technology perspective. I said, I got to have you on schedule, didn't work out. But fortunately, we traveled all the way from North America to have this conversation. >> Yep. >> First off, talk to me about strawberries. I'm a big strawberry fan and I went to the website.
I'm like, "Oh, organic strawberries. " What does Nature Farms, Nature Fresh Farms, what do you guys grow? >> So we grow bell peppers, tomatoes, cucumbers. And starting next year, 2023 season, we're actually going to start growing 45 acres of organic strawberries. So your favorite, so it's actually naturally grown and ready to go. We do both conventional peppers, tomatoes, and cucumbers, and actually, organic of the same variety. >> So the terms organic and technology, typically, don't... they don't seem to compliment each other.
>> Yep. >> Talk to me about how you're using technology to help you with your organic mission. >> Yep, so one of the things that we're doing is, we actually don't just take IoT to the small frame. We actually take it right to the extreme. When we look at a plant, we actually monitor a plant and the life of that plant. Everything from what sun goes in to how much nutrients goes in, how much it absorbs, how much it doesn't absorb, do we need to give it more light, do we need to know how much it's doing?
We absorb all that and we actually, from an IT size commoditize that. 5 million plants. I had to do some statistics on it. 5 million plants. Each plant at the edge, creates about 25 gigs of data for its life cycle. And then we store that at the edge, process that in real time, bring back just the information we need to the core. And we analyze that using a algorithms and stuff like that to make optimize the growth. >> So, and this is why I get excited when you stop by the booth, those numbers are mind bended.
>> Yep. >> 25 gigs of data per plant. 2 million plants? 5 on average. 5 million plants per season. >> Yep. >> Okay, we're going to stick a pin in it, and come back to the data question. But first I want to get to the value. You folks have been doing this for a while, does it make a difference? >> Yes, it does, for us it does. We're all about creating the quality. We want to give you a quality pepper, tomato, that you can take home to your house.
Create the best taste, create the longevity that you want at the house. It's one of our missions we've done. And by allowing us to control the environment, and not only controls the growth of the plant, but it allows us to maximize it. We actually have an R&D facility that every year, continually tries different variations of peppers. Like, we'll have 50 to 80 different varieties of peppers that look the same to you and me, and I didn't know this until I got there but they're actually slightly different.
And we test them to see which one's going to survive better, which one's going to test taste better. We have even taste testing groups, within our company to see which one does better. >> I would volunteer for that job. I would definitely volunteer for that job. >> Well, I did too, but I failed the taste test. (host laughing) because I couldn't get the acidity or the taste or you have to get certain things right. >> It tasted good, right.
>> Yeah, yeah. >> It all taste- >> Yeah, it all tastes good. >> These tastes great, that tastes great. " >> So, I'm fascinated by a few different aspects of this. First, I don't think we mentioned it, this is in the controlled environment, it's a greenhouse. >> Yep. >> So you can turn up down lights, you can adjust for pieces of data. You have to have data scientists. >> Yep. >> Managing at these numbers. Talk to me about the data science part of the business.
>> Yup, all of our data scientists, I call 'em grower scientists, most of 'em start off as a grower or working as a grower. And I'm now kind of migrated to working at the computer. >> So domain experts, >> Yep. >> They know the process of growing. Now they're putting data science behind that process. >> Yep. And helping them improve the yield. Every year, we try to improve our yield by 3% to 4%. And it's something that we're really say, I'd like to say we achieve not all the time, but as much as we can, we get that increase.
So that allows us to grow not just like a typical farmer, one plant a year. We grow tomatoes all year long. So our plants don't harvest once, we'll harvest all year. A tomato vine will get to be about 50 feet long, before we're done with it. And that's just because it's time to grow up. And then we plant one right next to it and start picking right away. >> So extremely efficient. You have growers who understand soil, soil nutrients, et cetera.
They become data scientists. You're getting all of this IoT data off of the plants. From I'm assuming nutrients kind, like what is the data that the plants get off? Like what do you track? >> So we collect everything from CO2 levels, nutrient in, nutrient out, how much water they're absorbing, how much light they get? The weight of the vine, pictures of it. We actually record pictures to know what it is. We're looking into different, other vendors that we now are using to look at when to pollinate.
When do we pollinate the flowers? When do we pick, how do we count it, do we got too many? Like we count how many in a cluster of tomatoes is there six, is there four there? Are we too heavy, should we take two off to help grow it? Like when you used to prune trees and farms like that, you got to take one off 'cause next year you got too many. And we look at all that to optimize the growth to get that maximum efficiency.
>> All right. So let's get to the infrastructure technology. 5 million plants, 25 gigs of data. Obviously, all of that data isn't coming back to a centralized place- >> No. >> that's just the network cabling along will overtake the infrastructure. >> Yeah. So we do a lot of edge computing right at the edge. So we actually do it because we can't bring it all back to the core. >> Right. >> And we want to be able to react instantaneous, 'cause the plant that's hungry or thirsty for water, if it waits minutes, it's too long.
Because there is a perfect time to water a plant. And it's something I didn't learn until a couple years ago, that there actually is a time when you've did, just like yourself, if you drink water all morning, well you didn't really get as much out of it. But it's that perfect moment. And that's one of the things that our data scientists, our grower scientists look for. Is that perfect moment to water the plant each day and how much to give them, to give them that perfect boost.
>> So are these systems tied into, are these edge systems tied into like irrigation, et cetera? >> Yep. >> So if the plant needs water there's not a grower that has to go out and physically with a sprinkler bucket and pour water. >> Nope, our entire system is automated. Like everything is done. Our growers can actually sit on their phone at home and basically, irrigate the plants. Or well, it automatically has a set schedule that they've learned, but it actually has an AI algorithm to help them optimally know when to do it.
So that way they know. >> So how do you... Let's talk about the control and update when there's a... 5 million edge devices? >> Story's an interesting deployment, that's for sure. It's something that changes, when we deploy it we just continue are teaching our AI algorithm. So it's never really a big deployment. But it's more of a newly refined data set that we've now figured out. We deploy it out there. " >> So it's a constant learning algorithm.
>> Yeah, it's constantly learns and learn. It does some deep learning of what the plant does. We've done all kinds of AI algorithms for stuff like that. >> So from a physical infrastructure, this is probably the largest edge deployment I've ever seen. 5 million devices? 5 million plants. 5 million sensors, that's a different struggle. >> Yeah, oh yeah. We don't quite struggle all the time with it. If you'd be surprised, you know, the industry and especially, over in Europe has really grown.
Like Europe really is a start of greenhouses and they know how to make their sensors. I will give them that. We don't face much downtime with the sensor. " Most of the down actually comes from the infrastructure side of getting it to the edge device. >> Wow. >> So, yeah. >> So what data comes back to any type of core system, if any? >> So what we do is we parse the data and we don't take the repetitive data.
We look for the changes only. And then we condense that, and we do a lot of things. All pictures usually stay at the edge 'cause we don't really need those back into core. " >> And then from a connectivity perspective, are these network via some type of Wi-Fi , do they have built in Wi-Fi ? How does the data actually get off of the device? >> So it's a lot of combination of Wi-Fi , fiber and even 5G now. >> Oh, wow.
>> We're actually starting to deploy 5g, 'cause again, we're in rural communities. >> 200 acres. >> Yep. >> 200 new to would be 250. >> Yep. >> Close to 250 acres. >> Yep. So it's hard to get Wi-Fi to work in a greenhouse, because you think about you got these huge rows of plants, 16, 20 feet tall and it just won't penetrate you. >> Penetrates are some amazing challenges. So and I'm a networking guy by heart, so it is hard for me not to resist the network conversation.
From a 5G perspective, is this vendor managed 5G or have you folks deploy your 5G network? >> It's a vendor managed one. We actually worked with Rogers in Canada and they've actually deployed a tower in our property to help increase the connectivity that we had and to give us 5G. So it was really great upgrade for us and it was great reception for them. >> So I'm going to do the math offline. 5 million devices. That's a lot of data at the edge.
This is a controlled environment. That much technology gear has to create heat. How do you compensate for kind of that IoT sensor heat coming off the devices for the plants, the plant health? >> So we are actually everything we have out there is rugged in design. >> Right. >> So we don't have to really worry about the heat. Yes, they do get hot, they get warm, but everything is built to be edge technology. Like everything is hardened, rugged.
Everything from the rugged laptops to the rugged sensors. Everything is designed to handle it. We have a lot of cooled roofs and one thing we are good at is absorbing heat. So we have actually looked at ways to funnel our heat out of our server room and kind of heat the water that heats greenhouse. We haven't done it successfully, yet but something we keep playing with. >> Yeah. I keep seeing these projects that are trying to recycle data center heat into some...
like a there's a YouTuber that has a rack in this house and he's trying to use it to heat his pool. This is a surprisingly difficult problem. >> Oh, yeah. Yep. >> Like you would think, wow, this this huge energy source, or this huge thing that's consuming energy. I'm organic and I'm trying to reuse some of this energy, et cetera. So how do you do that? So let's talk about that last thing, which is sustainability. >> Yep.
>> This is a lot of devices. You folks are actually doing a lot of great things. How do you power all of this technology? >> So powering being on the technology side or just power to give us energy. >> Powering to give you energy. >> Okay, so we use actually wood chips. We actually use wood chips, we have actually on our two ranges in Lington that use wood chips. We actually turned off the natural gas and we probably won't use it till Christmas.
We burn wood chips to heat our water that then, goes into there. The only reason we would turn on our generators right now in that area is because we actually also use the natural gas to heat the water, but we recycle the CO2 burn off and funnel that to the plants because CO2 is a natural stimulant for the plants. >> I was going to hit you with the way your burning wood though. >> Yeah, yeah. >> And, but you're growing plants.
>> Yep. >> And plants feed off of... I had to do a photosynthesis, a project with my 14-year-old, literally- >> Yep. >> Three or four days ago, and it hurt my head. But it's natural. You're kind of the perfect business to burn CO2. >> Yep, that's what we do. And if we don't burn CO2, we actually get liquid CO2 and inject that into the nutrients of them. And then we monitor that. So we, like I said, everything in that plant's life from the day it's planted to the day we rip it out and it goes to be compost it is done.
>> So Keith, this has been well worth the 12, 14-hour trip over to Barcelona to have this conversation, I looked forward to it. And you did not disappoint. >> Oh, thank you. (laughs) >> You bought the data points. ca I'm looking forward to somehow getting my hands on some of these strawberries after the 2023 season. I might even eat the tomatoes, I'm not a huge tomato fan, but I'm also a big fan of bell peppers and cucumbers. So, you know, one out one out of five ain't bad.
If you want to find out more about the CTO Advisor. com. If there's questions for Keith, that I did get to, you can DM me on Twitter. Twitter DMs are open. Until then, stay tuned for more coverage from VMWare Explore 2022 from the CTO Advisor. Borrowing something from my friends at the Kube, The Kube is the leaders in technology coverage, but the CTO Advisor is the leader in talking to enterprise architects and CTOs about the actual technology. Stay tuned for more coverage.
