Sign-up for the latest Lookout news and threat research
With more devices than humans in our world today, the amount of data being generated is higher than ever, with no sign of slowing down. Enter: edge computing. In this episode, host Hank Schless is joined by Said Ouissal, Founder and CEO at ZEDEDA, to discuss edge computing — from its business applications to the challenges of securing the technology and everything in between.
Hank Schless 00:08
Hi, everyone, welcome to security soapbox. I'm your host, Hank Schluss. And today, we're going to be talking about all things edge computing, which I'm particularly excited about as a relative novice to this topic. From autonomous vehicles to bot factories, the amount of data that devices generate in our world is arguably higher than ever before and keeps increasing every day. And as this growth continues to increase, and the growth of sensors, there's really no sign in any reduction in the flow of data. And that's where edge computing comes in. So I'm very excited to be joined by a true expert in edge computing, to help us learn more about it. aside, we Saul is the founder and CEO of zudena. And he's here to talk about this really exciting new field from its business applications to the challenges of securing the technology and and everything in between. Now, site is really dedicated his career to building out internet infrastructure. And he's originally from the Netherlands, with roots in Morocco, and is based in Silicon Valley. So truly a global individual. He also speaks five languages, which that's pretty impressive I speak to but that's pretty awesome. And He's authored two patents. So say, very excited to have you on here. Welcome to the show.
Said Ouissal 01:19
Great, thank you for having me here, Hank.
Hank Schless 01:21
So to kind of kick things off, could you tell us a little bit about yourself lots of interesting little factoids here, but also how you got into this industry and into edge computing.
Said Ouissal 01:30
Yeah, sure. As you said, I've been in infrastructure for all my career. So I've been working for companies and networking storage service providers, always building out the foundational layers for the internet. If you kind of quickly look back at my career, my first jobs are all about connecting continents and service providers. So we're building out as big internet backbones connecting US, Europe, Asia, Africa, etc. And then I worked for welfare companies that were doing everything about connecting home. So broadband was kicking in, and we would get fast internet, always on internet to the homes. Then I ended up working for a small company called Ericsson, which was the leader in mobile network communications. And we did 4g, we rolled out 4g, we brought, you know, basically fast internet to all our pockets. And what basically, I saw was like, with clouds emerging at that time, and doing a lot of work there as well, that it was very clear to me that the next step is all going to be about IoT connecting devices to the network connecting machines connecting vehicles. And that next evolution is driving a complete new architecture of the internet's and the cloud. And that's what we're doing here. It's a data. Very
Hank Schless 02:30
cool, yeah, IoT is something that's always sort of been a little bit like the distant cousin to mobile as well, right, kind of thinking about things in that light, but to sort of focus on edge itself. Can you tell us a little bit about it? And really, its history? Because I'm sort of interested to see, you know, how do we get to this point?
Said Ouissal 02:47
Yeah, no, it's great question. So I mean, if you look at mobile, and as you said, they are distant cousins, very related. And mobile has always been about consuming data, right. So basically, we are consuming a lot of data, we obviously contribute some data, we upload our pictures, and all of that, podcasts and whatever. But overall, we are consuming significant amounts of data from the network. And we have actually built on networks asymmetrical to support that. So if you look at 4g, or 5g or any network, it downloads very fast, and it uploads very, very slow. Now, if you shift to IoT, the distant cousin, it's actually the reverse IoT devices don't consume data, they generate data. A camera is a great example. It pushes a video into the network or audio or vibration sensors, or whatever the IoT device is. And this is where we saw five years ago, when we started to Dida, sort of like the architectures were built to do CDs and distributed content to phones and all that, that's great. But that's not going to work for the IT world, it's going to be the complete reverse, we're going to connect a lot of devices, they're going to upload a lot of data into the network. And five years ago, everybody was saying, Well, we're just going to upload all that data to the cloud. And this is where we were really confused. As you know, networking internet, folks were like, This is never going to work. There's a lot of data coming online, on Netflix that built for downloading data and uploading data. And not all data deserves to be uploaded all the way to the cloud. But you got to process it, you got to figure out what is the useful bed, so you got to run AI on the data or whatever else. And this is where we realized that the cloud would have to be pushed all the way to the edge of the network. There's cloud native concepts, cloud native applications, Software as a Service, all of that would have to be created, instantiated very close to these IoT devices. And that was the idea that we said, well, that's going to happen, people, we're going to need a way to make that possible. And we want to do that for them.
Hank Schless 04:28
That's, that's pretty interesting. It takes an intelligent mind to be able to do that. And in the midst of sort of the boom of the cloud, be able to think what would be next beyond the cloud. And you know, in the examples that you gave in sort of with cameras and things like that, I think when a lot of people think about IoT, or kind of thinking about the edge a little bit with sort of what little is broadly known about it, maybe they think about it on a pretty basic level. But in your experience, what are really some of the major use cases or major industries that are taking advantage of edge computing, kind of beyond the basic security camera example or something like that. Where's it really be applied more broadly?
Said Ouissal 05:01
Yeah. So the concept in itself, we call data gravity, right. So the data gravity is shifting to the edge, the application will follow the data gravity and needs to run at the edge. So if you think about it, from a conceptual point of view, where you want to go for is markets that have a big imbalance between the amount of data gets created locally, any amount of bandwidth available to the cloud, to upload that data to the cloud. So that's really the markets we decided to go first after. So a great example is oil and gas markets, oil and gas, you know, we go and drill wells. And renewable energy, by the way, is another example in that transition that's happening in that market. But we go drill wells, and we put solar farms somewhere in the middle of nowhere, there's very little connectivity, yet, there's a lot of data being created or gathered in those environments. So they are going to be primed as the first verticals that will run into this problem like well, how do I process and analyze all that data locally in a cloud native way? So those are the markets that we went into first. And we were right, actually, our first customers were oil and gas customers with a strong vertical for us, we just closed a series B round recently, and we had several strategic investors that participated in Iran, and Chevron was one of them, right, so that kind of tells you that they think this is strategic as well, from their point of view. So I think that the use cases there are analyzing wells, helping with drilling, for instance, monitoring wells, once they are, you know, dug and created, they're not always producing, they closed them down, but you need to monitor the pumps and keep the pressure on the wells. So these are all example use cases that we've seen customers go in and deploy, to shift
Hank Schless 06:29
a little bit. There's this, I'm interested to hear your opinion on the relationship between edge and cloud. Because, you know, I think that when people were going from on prem to cloud, it was like, Okay, so where how are these two going to coexist is one gonna take over the other, even still, now, with Cloud being so dominant? You know, you look at a lot of these industries, maybe like financial services that have such heavy data compliance, or privacy, or whatever it might be laws that require them to keep a lot on prem. And that process of sort of web enabling the on prem assets to make it look a little more like cloud, but still on prem. And they're also trying to implement cloud for other parts of their business that's kind of creating this very hybrid world, what would you say is sort of, like, take all that and kind of contextualize it around edge? And Cloud? You know, do you think that they're going to kind of happily coexist for a long time? Do you think that will edge sort of be the dominant one, do you think in a few years, what's sort of your perspective on that?
Said Ouissal 07:24
Yeah, that's a great question. I mean, I think a couple of points, I think, first of all, Cloud has had a profound impact on our industries, right. And, you know, the largest companies in the world today are cloud companies at the core, right. And, frankly, we're still getting started on the cloud journey, most major enterprises are just starting to move more and more workloads and applications to the cloud and enable new service in the cloud. So we're very early days in the cloud. So Cloud is going to be big. So I think that's kind of the first piece of it. I personally believe and that's what I'm doing is the D days, I think edge will be bigger than cloud. It's not going to happen tomorrow, it may take a little bit of time. But if you think about the sheer scale of billions of devices, and you know, the number of locations in the world, and we'll have more devices in humans, you know, that scale is just massive, right? And you think about it, but it's gonna take time, just like Cloud, we're 15 years in with cloud, and we're still getting started. And probably the edge is where cloud was 15 years ago, 12 years ago. So that's kind of the first way how I think about it. So this is a long journey. When we started and the ideas of edge computing started, a lot of people said, well, the edge will replace the cloud, right? We don't need to call it anymore, we can do everything in the edge. I think that's not the way it's going to work out. The cloud has benefits, centralization of resources, centralization of storage, data, analytics, learning, those are tangible benefits of enterprise are taking advantage of today. And they will continue to do that. So I think edge and cloud will coexist, even though as maybe they're going to cloud, it's going to continue, as we call it. And then I think the next thing is there's after edge sort of took off in the last few years, a lot of companies that are not cloud companies started calling themselves edge companies, which I understand because what do you call yourself? If you're not a cloud company? Okay, well, let's call ourselves an edge company. But the reality is, is there's data centers that are moving to the cloud, right applications that are running data centers that are eaten entirely being migrated to the cloud, or rewritten to run in the cloud and cloud native way, regardless, the combination of the two, that trend, that inward movement is very different than what we're seeing our customers do, which is they're connecting devices to the network that are generating data. And they need to now push processing from the cloud to the edge of the network, even further than the number of data centers, we had, let's say we have 1000s of data centers in the world 10,000, we're going to millions of locations of processing this data. So that's sort of the confusing trends that are going on within that which sometimes people, you know, from a simplicity point of view, say, hey, everything is gonna go to the edge or the edge will kill the cloud, or we are now an edge company. That's all great. But I think when we look at customers, it's a lot more of a nuanced view there. Now, what I think is most interesting about this pushing up the cloud to the edge, not the consumption model of the cloud. But the development model of the cloud. The way you can build apps in the cloud is 10 times faster or 10 times easier than you The old embedded software development that, you know, I grew up learning when I was younger. And basically, we have modern languages, we have a lot of tools, a lot of open source and a developer today, they can like create an amazing app with very little effort and resources because of that. And that capability is what customers need at the edge. And I think that's really how we thinking about is like, how do we empower these developers to build apps that exist between cloud and edge in the same ease as they do today? Just in the cloud?
Hank Schless 10:28
Interesting. And one thing that always comes up, especially nowadays, when people talk about this more kind of DevOps, see cycle that everyone's embracing, is where security fits into that, right? They're thinking about how do you integrate security into the DevOps workflows into that cycle, without slowing down? What's usually a pretty efficient, you know, two week process with patch Tuesdays, or whatever it may be? So when it comes to security in edge computing, what are some of the concerns there? What are you guys seeing in that regard?
Said Ouissal 10:56
Paramount concern, because if you think about it, we're going often in critical infrastructure, like electric vehicles, or the grid, or we talked about oil and gas. And imagine when in IT security is mostly data breach, you know, losing information about people in OT world OPERS technology world, these industrial worlds, security could be lethal effects. Imagine you have a safety system that makes sure it shuts down robots, if humans walk into specific area where you know, things move around, then you don't want a person to be there. Imagine you hack that system, and somebody you know, is relying that system to kick in, or we would normally kick in and it wouldn't, it could be lethal, right? Or imagined shutting down to great, like we've seen in the news that there's a lot of concerns about this in general. So security is of a complete different dimension, and a complete different aspect. And, and what we saw early on when we started the company is not only do we have to bring that ease of use and new way to deploy apps at the edge the same way how you do it in the cloud. But we have to do that with security from the ground up built in. And we spent a lot of time actually as a company in the first few years to build a super secure zero trust architecture, as we call it, that I think without we would not have had the customers we have today. And we've gone through each of them through significant security reviews, they don't easily put cloud connected or network connected systems in these LT environments, and you have to really hit a higher bar than you would normally have to predict.
Hank Schless 12:16
Yeah, completely. And especially in that kind of oil gas energy sector. In particular, we've seen the same thing on our end. And
Said Ouissal 12:22
the network security is one part. But I want to also highlight one other part was the physical security because you're not putting tiny compute nodes out there. They're sitting in physically insecure location. So we had to build features not only to protect from network attacks, and all that stuff, but also what if somebody steals the actual device or tries to use a USB stick to insert malware in a device that then could install and then from there, use the device as a jump post into the network or further fine attack surfaces. So the physical security, we have to do it, frankly, the only industry that has solved that today is the mobile phone industry. So talking about your earlier distant cousins, we actually use technologies like you find on Apple or Android, like, secure and clap. And you know the way how your biometrics lock up your keys, and we use those in our operating system to provide that capability there. And we always assume that the node will be compromised. That has always been our we're assuming it's going to happen. And what are all the things we can put into a product to make it really, really hard for a hacker or attacker to exploit the note?
Hank Schless 13:22
Interesting. So that's really cool. I think people sometimes forget the physical side of it sort of in this very cloud world that we live in. So to throw a little bit of an industry use case at you here. Let's say you're a manufacturing company, and you want to move to the edge. What does that journey look like for the company? And what are the things that they might look out for?
Said Ouissal 13:41
So I mean, I think the first and foremost is the business value, right? Why go to the edge? Yeah, edge computing is cool. But is there a real business value? And what we typically see it starts first and foremost with agility, all these manufacturing example you're giving they want agility in their business. They want to know faster and quicker what's going on in their environment, what's working well, what's not working? Well, where can they optimize and reduce costs, or improve margins or improve revenue that's constantly on their mind. So that agility is kind of step one. The way to do that is through data. And so the more data you can collect, the more data you can analyze, the faster you can make these decisions, and the faster you get this agility in your business. So that's sort of what it usually starts with. So there's usually a data science and a data approach where customers go, like, we want to get this level of insight in our production. And all we want to know exactly within an hour of when something gets produced want to know what got produced, and how much it costs to produce it and where it is, and things like that. So that's where it starts with just in time, and real tightness of the business. And then from there, they start building a strategy of like, where's that data today? Some of the data may be already collected, but it's not connected. Like they may have an A machine that is today, you know, recording a lot of data through a software application called the SCADA application, but it's all stored locally on industrial PC somewhere, not analyzed. So they go like Oh, can we go in and add edge computing to that system and base sake we start processing data. Sometimes the data does not get collected today because vibration if you think about predictive maintenance and things like that, so they actually augment the device or production line with additional sensors. So they put like Bluetooth sensors or whatever else. And then so basically start seeing the sensor fusion idea of the data to have to day to day to day need. And then there's an edge computer at server that gets deployed, installed very close to production line, a production machine that starts collecting all that data. And then through our architecture, they can now start pushing software applications to continue to refine that data. If data is the new oil, then edge computing is the refinery of the oil, right? So they stopped and then basically, through software improve more and more and they go like, right now we see this, can we push out a new software version that now allows us to correlate this data with that data. And then once I have that working on a single machine, let's say they have 1000s of these machines in the world at 10,000, then they can with our system, push that same software to all of those nodes that are running around the world, and basically scale up that efficiency or that agility to the entire production. And that's how we've been going at it. So it's a very nice process and steps, we usually get involved one day, I've done the pilots, they've established the business value, they've seen that this has actual impact to the bottom line, and then they start figuring out like, Well, how do we take this and extend it to the rest of our environments?
Hank Schless 16:24
That's pretty cool. And, and thinking about the scale of all this, too. I mean, you sort of said before having more devices than people, you start to think about how wide this can reach and is reaching currently, and the scale that it will reach in the future. You know, one thing you mentioned there was being able to push apps and things like that. One thing that is top of mind, especially these days, whether it's on mobile, whether it's with a SAS app, whatever it is, is being able to patch and update apps, right, you know, patch security issues, there's a new vulnerability found in whatever app, whatever it may be. So on the edge, you're talking about, you know, 10s, you know, hundreds of 1000s of endpoints, you know, does that sort of patching capability that process sort of work the same as it would on a cloud on cloud infrastructure? Is there sort of a nuance to it with edge?
Said Ouissal 17:07
Yeah, no, we want it to be the exact same as the cloud. So the way people are doing it, until edge came around was something called over the air updates, OTA, and OTA was really built as an after you roll out a product or some embedded system, and you realize you have a bug, you need to go fix the bug, and you'll then roll out over the arrow. And then if you didn't fix the bug, then you have to do another over the air updates to fix the wrong software you rolled out through the first OTA. So you're always rolling forward, as we call it. And it's not really intended to add a lot more new functionality to the software as it is to further basically fix bugs and make the software work according to the original idea. That's the embedded way of thinking, right? We don't think that's the right way. That's the wrong way. And then let's compare that to the cloud way. If you're a cloud company today, in regards to your Facebook, or your Salesforce or whatever, you're pushing out constantly new software capability in your production environments, with one click, you upgrade all your services in the cloud. Or you can do that multiple times a week if you need to. And when something goes back, you don't roll forward, your roll back, you say, Okay, let's go back to the version of software that we knew was working. And it's very easy to rollback in the follow up because you actually maintain a copy of the old software and you just spin it up, and you're good to go. That is what we were bringing to the edge with our solution. So basically, the ability to continuously deploy run software as a service at the edge, manage it as if it's running in the cloud, and giving customers that same agility they used in the cloud for the software development and software rollouts, giving that to embedded devices, or machines, or vehicles or anything else. And that's really core to that, again, agility concept that the business is looking to accomplish.
Hank Schless 18:45
Interesting. So we're kind of starting to come up on time here. But say one thing I'd be really curious about is what do you think the future has in store for the edge? You obviously have made a good call five or six years ago, when you decided to start this company? And what would be important. So what do you think, is next for the industry itself?
Said Ouissal 19:04
A couple of topics. I mean, I think first and foremost, the advent of AI and edge, they're very natural fit to each other. Right? So we have actually several customers today that are using us to deploy edge AI models like inference at the edge learning in the cloud, but infer at the edge. So that's a very common model. And you know, we have to figure out how to do that in a hardware accelerated way using GPUs. And I mean, a lot of the problems people do when they do AI in the cloud, again, they want to do it had the edge, how do you give them the same experience? So that I think is an absolute super exciting trend? Because while we want to write all that software, we also need to leverage the power of AI to help us figure out what are the relevant data patterns relevant, this is the pattern of a machine that's going to fail in two weeks, right? If you have that kind of an AI model, figuring that out for you and continuously improving, the efficiency of that model becomes better and therefore your business becomes better. The other piece that I think we're seeing is we're going to move more and more from collecting and analyzing data to actually respond Knee to date. So when things happen, self tuning, self regulating systems, this is a tricky thing in the industrial world, because control is considered super critical. I mean, again, when you control a robot, you can do a lot of things with it, and not always good thing. So people are very sensitive about that. But people are realizing that that's where it's going to have to go. So how do you build the parameters and security around that you start controlling systems based on data generated and analyzed. And that in itself is a stepping stone, I think to the next big thing, which is autonomous cooperating environments, like robots working with each other in a factory line or vehicle, automotive vehicles, autonomous vehicles communicate with each other through V to X, or you know, those type of technology. So I think that is really what I think the edge will lead to is that autonomous world, that interconnected world. And of course, security is going to be essential, right? I mean, you don't want a fleet of cars that need to be controlled or turned into a weapon. And you want to make sure that these things are very restricted and cannot cause harm. So we think this will take some time. But that's really what's in store. And you know, a lot of the science fiction TV shows we see, I think will be enabled through edge and flying the right way. Very
Hank Schless 21:11
cool. Yeah, that's one of those things. I feel like we could do a whole nother, you know, half hour here on but I think that's the time we have for today. So say thank you so much for taking the time to join us today. If people want to learn more about you, or as a data, where can they go? Where can they find you?
Said Ouissal 21:24
Yeah, obviously our website is data.com. We've got LinkedIn, Twitter feeds, different social media, we also a founding member of the Linux Foundation edge computing organization. So these guys are focused about open source architectures, part of a software is open source. So that's another great source as well for developers or people that are really trying to understand the nuts and bolts. But yeah, there's a couple of different ways and looking forward to engagement. Anybody interested?
Hank Schless 21:45
Awesome. Yeah, just for the folks listening the DITA ze de D. D. A, just to make sure you get the right one side. It's been great to chat with you. Thank you so much. If you're interested in finding more interesting security and technology topics, be sure to subscribe to the security soapbox podcast. Visit lookout.com/blog to learn more. And as always, follow us on LinkedIn and Twitter at Lookout and until next time, thanks for joining on your host HLS talk to you soon.