An EPIC "Dream Machine" for Ignition Edge (Heads in the Cloud Podcast)
Listen to 4IR Solutions CTO Joseph Dolivo and CEO James Burnand talk about Joseph's guest blog on Opto 22's blog "The Art of the Possible: An EPIC "Dream Machine" for Ignition Edge".
All right, welcome to another episode here of Heads in the Cloud here with James Burnand and Joe Dolivo, where we are talking about IIOT Industry 4.0, as well as transforming your business to be ready for the future. For this topic, I'm so excited because Joe worked really hard writing an article for Opto 22 for their blog, The Art of the Possible, and we kind of want to dive in a bit more about that. It was the Art of the Possible and EPIC Dream Machine for Ignition Edge, and I think that James and Joe have some interesting insights in terms of how you can use that to your advantage, as well as how that would help in terms of moving your business along to being future proofed. So James and Joe, you guys want to go ahead and jump on into it?
Sure. Yeah, I'll as well commend you, Joe, on the well written article. I think the Opto product in particular does speak volumes for itself, and I know we've very selfishly chosen it because it solves a lot of the problems that our end users have with how to bring ignition into the plant when they're using a cloud-based infrastructure or cloud-based architecture. Do you want to talk at all about the wonderful things you wrote about?
Totally. Well, it was a good opportunity, I think, to sort look at the product in the space of the edge and edge computing and edge devices. And for us having cloud offerings and manufacturing, you need to have something that's living on site until we have 5G cells everywhere. Maybe some day in the future. It'll probably be 7G by that point before it's actually practical.
I was going to say, I probably won't get it. I'll have my tinfoil hat on so the signals won't get through.
Exactly. So there's always a need to have something on the edge that's going to be doing buffering. And in the case of ignition, local control fallback. So the edge devices are... There's a whole bunch of them. Ignition actually has a program called Ignition Onboard where they will effectively certify certain devices for using Ignition and Ignition Edge. Opto happened to have one of those solutions. And one of the big differentiators, and I'm going to talk more about this in the blog, but definitely a big differentiator is that it's a true PLC. So not only are you putting an industrial PC there that you're running Ignition on and maybe other software, but it's a PLC and it can communicate with field devices as well. So you can have kind of a one stop shop, if you will, for doing all of your local stuff.
And it just goes along with that theme of trying to reduce the footprint of this equipment that you have in the plant that somebody's got to manage and deal with upgrades and updates and all the stuff that we kind of talk about in the context of FactoryStack.
It's funny. You called it a dream machine for FactoryStack. I actually think it's a dream machine for OEMs. I know OEMs maybe don't get as much attention on our kind of blogs and that kind of stuff, but if you think about the concept of having to have some sort of onboard controls that's running your piece of equipment, but then also needing to host an HMI and also needing to and wanting to be able to offer your clients additional value by being able to do things like remotely monitor alarms and manage your systems and benchmark their systems versus the rest of your install base for the cost of a PLC and actually cheaper than a lot of larger manufacturer PLCs, you can get PLC functions.
You can hang remote IO off of it as well, as you can communicate to the cloud. You can do all sorts of advanced features using Ignition, but also Node-RED is on there. You can run CodeAssist, you can run a VPN client off of there. It's got two network cards on it, a firewall. Joe, I probably missed a bunch if you want to fill in the blanks, but I guess I look at it from a manufacturer's perspective. It's a really great way to very effectively connect in to the cloud and to make that edge piece work. But for folks that are actually deploying systems as OEMs, I think it's also a dream machine for them as well.
Oh, it totally is. And I think it also sort of walks a fine line between being open and being too open, which is something when you talk about security and all of that, that that becomes a discussion. But I mean, it's an industrial PC. It runs Linux. And to Opto's credit, it's an opt-in process you can go through if you want to get a license to get shell access to it, where you can pretty much do anything, limits the warranty and support and things like that if you go ahead and do that. But you can use it like an industrial PC in that case. But by default, it's very much locked down and they take kind of a batteries included approach where it's got the software that James is talking about, so you can do most of the stuff that you'd want to do, and it's supported by Opto, and those applications are basically baked, including Ignition, are baked into the firmware.
So you're not just doing these upgrades willy-nilly where you might break some compatibility there. You're going to flash a new upgrade or new firmware version, which is going to include all these things packaged together, which has been already tested and kind of validated to work together. So it's just a really nice package for doing all of that without you having to do some of that yourself with any old industrial PC that you can get on the market.
Yeah, and not to knock industrial PCs, I think there's definitely a lot of use cases when you're looking at higher bandwidth or needing storage or wanting to run databases where maybe the EPIC isn't always the best fit. But certainly a lot of the use cases we've seen things like running OEE systems centrally and doing all of your data collection and buffering at the edge, it's really a perfect fit for something like that. Maybe we want to transition from that though and talk a little bit more about, well, what do you do if I need more than what a simple Edge device... I don't want to call it simple, [inaudible 00:06:05] justice, but a purpose built Edge device can do to something that can actually run full workloads.
Yeah, for sure. I mean, even take a step back, even within the groov products line, you've got the groov RIO, which doesn't have the nice screen, which is another really nice thing that the EPIC comes with, but it basically will just let you do IO communication and it's actually also can have Ignition Edge built in. So that's for the lower end when you're just basically interacting with devices. You have the groov EPIC, which has more of these industrial PC functions as well. And with the newer version, the PR2 module, that's actually powerful enough to run Ignition 8, the full-blown Ignition 8 and everything that comes along with that, like writing scripts and connecting to devices. That gets you pretty far. But to James's point, then there's the point where you want to start supporting other software.
You want to have databases that are running locally. You have other software that may or may not be containerized, or maybe it's packaged up as a virtual machine and you want to run that, and that's where we start looking at the edge as sort of this component within a hybrid cloud architecture. And there's more applications, there's more stuff that you need to be able to run locally, and there's a lot of options out there. Some of them have been around for a very long time from companies like Nutanix. Some of them have a common virtualization hypervisor platform like VMware we're running on top of them. And then the cloud vendors as well have their own solutions that also kind of ride on top of that, but those are going to be for larger workloads, but still consolidated in terms of having everything kind of in one place and then having, in some cases, a subscription or a lease model that you can get these from a vendor to support. So I know you've looked at a number of these, James. What are some of the ones that you're familiar with?
Yeah, so it's neat. So disclaimer is that they're not available in all regions yet, so you do have to be pretty careful, especially if you're not in North America, to make sure that if you're looking at these things, that the hardware as a service model being offered is something that is available in your region. But the big three, AWS, GCP, and Azure all have basically edge offerings that you can lease from them, different costs, different capacities. AWS is called the Snowball, and it was originally kind of built and designed as a data capturing device that you would mail back so you could transport that information back into your cloud instance. What they realized was people didn't want to give it back because it was pretty useful to have that compute capability on the edge. They have different models with graphics capability for doing image processing.
And in the industrial use cases, I mean, they're capable of running a Kubernetes cluster and virtual machines, so you can do quite a bit with these. Again, their form factor is designed to be relatively useful in an industrial environment. In some cases, for example, the Microsoft version does go in a standard rack, but all of them kind of do similar things in that they'll allow for you to, for all intents and purposes, carve out a little bit of cloud down inside of your factory and then be able to run that using services that you deploy through your cloud console, but operate onprem. And if you really are able to handle things like internet disconnections and low latency application requirements, things like running batch servers and data collection and that kind of stuff, while still getting the benefits of using cloud-based tools for things like monitoring and disaster recovery and management, as well as, in our case, we operate a orchestration engine on there. So there's some self-healing capabilities of the applications themselves when they're deployed in this kind of an architecture.
The other two, just for reference, so Google has what's called the Edge Appliance, and then Microsoft has the Azure Stack Edge. These are the single unit, purpose-built, edge hybrid devices that you can lease from each one of those companies. And Joe, maybe if you want to talk a little bit about once you grow out of those, what the next biggest option is?
Yeah. Well, it's just interesting to look at those devices and see how they've kind of evolved from that use case of just kind of capturing data and then shipping it back to the cloud provider to dump it into their object storage model, kind of like the Snowball that James was mentioning. But beyond just doing local vision processing or inference of machine learning models and things like that, I think all three cloud vendors now have realized that, "Hey, there's actually some real compute power and we can take advantage of running any applications there locally." And so you can get pretty far with some of the resources that are there. They're comparable to maybe other, let's say, like a Dell rack server that you might be able to get yourself, but of course they're managed by the cloud providers.
When you need something bigger, then you can step up to larger offerings, which, again, the cloud providers have. So instead of a one U or two U rack unit, you can get a full blown rack. So AWS has Outposts, Google has a different distributed cloud product that is, again, more akin to having a full rack, and then Azure has partnerships with some other hardware vendors via their Azure Stack HCI or hyperconverged infrastructure family. So there's a few devices in that family as well that are just aimed for more powerful workloads beyond what you do on one of these little, "smaller" edge devices.
Yeah, I mean the options out there really make the case that you really shouldn't be installing standard hardware anymore, kind of the old-fashioned way, that taking advantage of using a hybrid model from a sensible use of cloud, even if you still plan to operate all of your actual workloads and your applications onprem, the benefits you get from the operation model is just tremendous. I also thought there's something pretty cool that AWS did a couple of years ago, is they have a product called the Snowmobile, which if you Google it is the most ridiculous thing I've seen in a while. It is a tractor trailer with a hundred petabytes of storage on it you can use to suck data out of whatever... that you need and then drive it back to the AWS data center to be ingested. So I have not found a use case for that yet, but I'm still looking.
Could be a part of our backup strategy for FactoryStack and PharmaStack.
That's great. I love it. So guys, this is about the end where we're going to wrap it up. I appreciate everyone for listening to the Heads in the Cloud podcast here with James Burnand and Joe Dolivo. And don't forget that if you wanted to actually check out that article, the Art of the Possible and EPIC Dream Machine for Edge Ignition from Opto 22, you want to make sure that you head on over to Opto 22 and their blog. So you can check it out right on their website, an EPIC Dream Machine for Ignition Edge. You're listening here on Heads in the Cloud. See you guys next week.
Heads in the Cloud, your source for future proofing your business.
Audio Podcast: https://spotifyanchor-web.app.link/e/gzLK3xcPMsb