Analyst Chat #104: Edge Computing

Welcome to the KuppingerCole Analyst Chat. I'm your host. My name is Matthias Reinwarth. I'm lead advisor and senior analyst with KuppingerCole Analysts. My guest today is Graham Williamson. He is senior analyst for KuppingerCole working in the APAC and Australia region. Hi Graham. Good to see you. Good to be talking to you again. Great to have you, and you brought a really interesting topic and the topic that I see in my advisory work on almost a daily basis coming up right now.

We want to talk about, about edge computing and how to do things around edge computing and how to do it properly and securely. But first of all, my first question to you, a definition, what is edge computing? Just to be concise and to be on the same page for all of us. That's a very good question because edge computing means different things to different people. So it sort of depends upon the industry sector you're talking about.

And the context within which has been used, the Genesis of edge computing was in the operational technology sphere, where organizations had large environments that were dedicated to our mission critical production. It could be collecting data from sensors. It could be managing a, a manufacturing automation operation that this is a situation where there's a lot of data traveling around a network that is mentioned critical. And so the question then comes as to, well, how does the company understand what's happening in that environment and how do they get data out of it?

And so this is where edge computing comes in because it says, okay, we don't need all of that data on the factory floor. Or what we just need to know is how many finished goods have hit the conveyor belt. So the edge, an edge computer then provides the supervisory capability so that the, the corporate management can understand what's happening. So they don't need to know how many times that drill press has operated, but they do need to know what sorts of production levels have been maintained.

So edge computing restricts access to the, the operational technology network to just that management communication that's required, which means that the, the, everything can happen in your team environment and not be like this. In many cases, it is pretty mission critical. So have you got administrative work that's mucking about in there that could delay communications and it could disrupt databases.

So, so you want to have this edge computing devices going to limit that it could be a PC, it could be a controller, it could be a number, a number of things can fit that category, but its purpose is to limit the access between the operational technology environment and the management environment.

Yeah, I would fully agree what I see in addition to, to what you just mentioned, that we see that this data, as you said, it's usually not needed for further purposes, but we see that some organizations already apply some big data technology to, to analyze the data in a very controlled and access controlled manner. So they push it into, into big data platforms, but really in a controlled manner and only for, for limited purposes. And then it's, again, this connected or really access controlled in a very, very strict manner.

So it's really also using that data for very specific purposes while they are not needed, for example, at the headquarters or somewhere else. So it's really controlled access when we, when we talk about use cases, you've mentioned one already.

So these, these, these OT environments, you mentioned that these are the starting points. What other use cases do you see or are there more around OTs? Right.

So I, I was invited to write a blog and as I was, as I was discussing the edge computing environment, you start to think of the other, other environments, like a vehicle like a, I've just, I've just actually purchased a new vehicle, a first one in 10 years. And I am boggled by how much has happened and how much computing is happening in a vehicle now. So if you consider what's happening throughout that vehicle from the, the engine, from the brakes, from the it's just about the environment, all of this now is, is being affected. All of this data is being collected.

Now, you don't want that information. Oh, and there's another, there's another interesting thing. Now that's happening in vehicles and that is the, the, the communication from the vehicle to roadside assets.

So the, in, in an automation environment, we need the vehicle to be able to communicate with the traffic light that is coming up to, and if the, if the car's got his left turn signal on, we want the traffic light to give me a green pen filter and these sorts of things are happening. So, so the amount of computing that's happening in a vehicle is just going to, so the question now is, well, how do you protect all of that?

And how do you make sure the it's not interfered with like really the only time that a human actually needs to, to, to, to get information from the vehicle is the service mechanic who needs to understand and run a diagnostics on the system, potentially a police investigator. Who's wanting to, to, to see how fast the vehicle was going, like when the accident occurred and things like this. So now we need an edge computing device that allows that communication to happen, but doesn't interfere with the mission critical activity that's happening, particularly when that vehicle's in motion.

So, so that, that to my mind, there's another very interesting use of edge computing. And if you're talking to a vehicle, a designer, somebody who's in that spite sphere, when they say it's computing, that's what they're talking about. So it's another parallel use of that term edge computing, Right? And it's not only about security, it's all also about privacy. So there's lots of very, very personal data involved. So really nobody has to know where I travel, unless there is a good reason to, to disclose this information.

When you talk about vehicles, I saw just as a quick side note, a very interesting edgy use case it's really freight tankers or freight freight ships or tankers that are loosely connected, but also out more or less a swimming data center, but only loosely connected. And they need to be managed as well. So these are also vehicles, large ones, but this is really something that organizations need to have access to during that travel of, to find out where they are, how the freight is, if, if delivery has worked out correctly.

So that is vehicles at a larger scale, but you've mentioned the car, the vehicle w when we look at our own home environments, there's lots of computing power right now there as well. So I need to not to speak up, say something like Sierra or Alexa, otherwise they wake up. So this is also something to look at, right? Absolutely. And that's one area that this whole edge computing is going to become so important. Like in my house, I've got solar panels on the roof and then I've got a solar water heater and things like that. It's pretty basic.

I must've made the noise mean that, you know, my, my solar panels generate all this, the DC, the recurrent, and then that gets through an inverter made into AC so that I can feed back into the grid and how all the equipment in my house. But then my TV takes that information and puts it back to DC. And so every time you do that, that's a of energy. And so I can see a future where most of the devices in the house is going to be DC.

And w the, there is going to be a need to go to AC for the grid, still working on AC. But then again, know if I'm feeding back into the grid. I want to do that only when it's, when it's advisable to do that.

When, when the spot price of electricity is high and it makes it worth my while to, to put power back, back into the grid, and then there's the electric vehicle or I'll have. And the electric vehicle is going to need a special charging mechanisms and ability to take the power and put it into the vehicle at different rates, depending upon the charge rates in the battery, there's an enormous amount of processing. There's going to be happening in our houses. And so the next question comes as well. How are we going to number one, protect that.

And number two, make sure that we don't put out all of this data on our internet. You know, we don't want all of that information going out on the internet. We want to make sure that our, our data costs minimized. And so we needed an edge edge device. It's going to be associated with our wifi router, but it's going to now control what we are allowed to come in and out. And then all of the other processing that's happening on my home network is not going to get interfered with. Right.

And I think we are not talking about something theoretical, the examples that you've mentioned, they are reality. They are here right now. And it's really growing area of, you just mentioned a few examples. There was much more about that. So really what would be the, the analyst's approach towards doing edge computing in each individual case, right. Is there a common denominator to say, okay, security plus costs how to do that, right. What would be an approach for doing that? Yeah.

So those are the two biggest security I'd say, you know, in some environments, safety is important to know if, if, if we've got a network where a low Natan latency is very important. You know, if, if an alarm goes off, we've got to take action right away. We don't want, you know, a couple of second responses, it's gotta be a millisecond response. So safety is an issue, but security is going to be a B, because we obviously do not want the data exfiltrating from our network. Like if we're doing a production data in, in a, in a manufacturing situation, that's pretty sensitive information.

We don't want our competitors to get access to that. If we talking about a vehicle, as you mentioned, we don't want people to know where we've been or, or to be able to track us in, in any way.

So again, this security is, is a very important situation. And obviously for our home, we need to be very careful about what information goes out on the internet. So those are the security issues we have. But as I mentioned, cost cost, particularly let's discuss the, the factory situation.

Again, the factory needs to communicate information back to the head office. Head office needs to get certain bits of information, but if you send everything out over that internet connection, that would be enormously expensive, even with 5g. And the beauty of 5g is you can buy just the data you can pay for just the data communication that you need. You don't have to, to buy a circuit.

And well, I can, you know, it's unlimited on that circuit. You're wasting money if you're doing that. So in the 5g space, we need to, we, we need to be much better at deciding how much data do I send, where am I going to send it? And then just paying for the amount of data that I sent. So let's take, for instance, a building. So our building has monitors all over it and equipment and on elevators that there needs to manage air conditioning systems that need to manage. You don't want that building, communicating all of that information across the internet.

You want that happening within the building information management system, and then the building information management system, just sending that few kilobits every five minutes, every 10 minutes or whatever is required to keep management happy, happy. And, and you, the dramatic decrease in the cost of, of, of communication.

If, if you do it intelligently, I remember reading about it's actually a retail situation. So the, the it's in the UK 10 stores where with a head office and in, in 2020 was the last was, was the first time that the company had seen a reduction in their communications, because they'd been able to do it intelligent your POS systems. Don't need to communicate everything back to head office, just the information that you need. So by again, using intelligent communication, we we're going to be able to reduce our costs.

And this brings up the other, another interesting thing is who is going to be at the forefront of providing the edge computing capability. Who's going to be designing the device that does that edge computing activity, who's going to program it. And where's that all going to come from. And I think there's a, there's, there's a green field of anybody that wants to get into this area.

Now, edge computing would be a good one to get into the telcos, I believe are going to be seeing themselves as being the major purveyor of capabilities in space as the, the actual amount of revenue that they receive from communicating data on a per gigabyte basis drops, they're going to need to move into other services. And this is an ideal one because obviously they understand 5g core capabilities, how to do the data, data, slicing network, slicing to minimize data usage and these sorts of things.

But I think, yeah, there's, there's, it's, it's another one of those areas that it's going to burgeon in the, in, in, in the very near future and become an area for, like, if you've got any kids coming out of school that wants to know what area to go into edge computing would be a good one. Yeah. I would fully agree as they are these various and very different use cases that you've mentioned, I think making or doing edge computing, right. Involves what you said, re reduction of actually the data that is really traveling and leveraging the processing power that is at on the edge.

And there is lots of processing power. So having the right combination of, of de-centralized processing and communicating the right information to those who needed that will be the real challenge. And that is true for the home environment. That's true for the mobile vehicle environment. And that is of course true for the, for the corporate and for the enterprise environment, with lots of OT going on in different locations and communicating only the important information while extending the security to these areas.

The, the, the corporate network has just become even more complex with adding edge, which is a challenge, but it's also comes with, with lots of promises as a, as a final note for the, for the home user, for the end user. Is that something that you would recommend just doing right now when there is already some of these edge computing devices in here, checking for security? I think, I think one area that's been solely mismanaged is this whole area of internet connection that the average resident has. Okay.

I mean, although most suppliers now have fixed it up, Wi-Fi routers that are being shipped with default passwords, you know, admin as the username admin as the password. No, that's got to stop and people need to be helped to understand how to go into the admin settings and make sure that the, the password is being used is, is one that they want to use in complex enough to make sure that the brute force attack wouldn't, wouldn't be able to detect it.

So, yeah, it's, from my point of view, we've been very slack at our security on our home devices, hopefully as we get better. And as the, as I mentioned, the home being more, more of a network being more and more integral to the what's happening within the home and more and more in need of, of protection in terms of the security aspect of their equipment, hopefully we're going to, we're going to see much, much better. We have.

I mean, I've got to admit over the last four or five years, there's been restrictions on, on companies. Most companies are having to make sure that that wifi routers now are shipped with a set with their own, their own password. That's known only to the recipient.

And, but again, with the recipient, being able to change that and knowing how to change it, that would just heighten the security that we have at the minute. It's very hard to limit the amount of communication that's available through your wifi Rafa. So I think that's one area that we're going to see a lot of work in, in this edge computing capability that our wifi rats will be able to provide this. All right. Thank you very much.

Also for that recommendation, of course, I would highly recommend for those who are interested in learning more about that to first of course, go to our website KuppingerCole dot com and find out more about edge computing just by typing in edge into our search engine that would lead you to the right places to look for. You've mentioned that block posts has that already been published, is that online so that we can link to it in the show notes?

Yeah, but I'll leave that with me and I'll get back to you on that and for us. Okay. That's great. And we will, we will really provide the audience once it is available with the information where to find that. So thank you very much, Graham, for being my guest today, that was really an interesting topic. And it's really something that is just out there right now. So we really need to take care of that and manage it properly. So thank you again for being my guest today. My pleasure. Thank you.