Harness the Power of Edge to Cloud Architecture – Webinar Recap

Inductive Automation invited our team to participate in their webinar series to talk about harnessing the power of edge to cloud architecture. You can watch the entire webinar here or read the recap below. 

Meet the speakers:

Travis Cox: I am the Chief Technology Evangelist at Inductive Automation. In my role, I help to educate and develop awareness of the Ignition platform and supporting technologies that are available today. 

Mike Walsh: I am a Senior Product Manager at OnLogic and I work with our customers and partners, like Inductive Automation, to make sure their product needs are being met and to ensure that the evolution of our products is delivering the necessary features and functionality.

Cole Wangsness: I am the Partnerships Leader at OnLogic and I work together with customers and our partners to build better solutions together so that we can deploy software and the hardware solutions together.

Webinar agenda

• Introduction to Ignition and OnLogic
• What is the edge?
• Edge benefits
• Hardware & software considerations
• Cloud
• Going forward

Intro to Ignition

Slide courtesy of Inductive Automation

Travis: Ignition is a universal industrial application platform. It’s basically an enablement platform that allows you to build HMI, SCADA, MES, and IIoT solutions. It is used by 44% of Fortune 500 companies and 57% of Fortune 100 companies across all industries around the world. It has an unlimited licensing model, so you get a server license that is completely unlimited in terms of tags, screens, clients, device connections, and more. 

Ignition is fully cross-platform to run on Linux, Windows, and Mac OS, and it’s perfect for the edge. It is built in IT standard technologies and uses modern technologies. It’s also a scalable server client architecture that is completely web based and web managed with designer and clients that you can launch anywhere.

Ignition is completely modular, so you can pick and choose the functions that you want and you can easily expand as you go forward. It also has a very rapid development environment that is easy to deploy out to clients. 

Background on OnLogic

Cole: OnLogic makes highly reliable, easy to install industrial and rugged computers. Our Ignition Line comes pre-loaded with Ignition Edge, but Ignition can be installed on any of our systems.

Since every project is unique, we offer a variety of options. 

• Industrial Computers
• Rugged Computers
• Panel PCs for HMIs
• Edge Servers

Everything is built to order and you can configure how you want it with the I/O that you need. Our objective is to make it easy for us to be your hardware partner so that you can focus on your software and your application. We offer a consultative sales model for anybody. You can work with a salesperson and for larger projects, you can work with a sales engineer to determine what you need. 

You can get a prototype assembled quickly in days instead of months, and when you are ready to deploy, we can provide additional services like pre-imaging systems, bios customization, or branding to really make it your own. We also offer lifecycle management. 

We were founded and headquartered in the US and we have a global reach including offices in the Netherlands and Taiwan. Our systems are built right here in Vermont and we also do manufacturing in the EU. When you work with a sales person or a support tech, you will most likely speak with a real person in your time zone, and that’s something we really value. 

What is the edge?

Travis: The edge refers to the edge of network nodes that collect data and transfer it to a central location. It also refers to functionality that is right there at the edge, such as a local HMI and even running machine learning tools or analytics at the edge.

It’s that compute that allows us to either move data or provide some local functionality that’s required. When we look at these critical systems that we have out there, we can’t rely on connections to the Internet. We have to have edge devices that are close to that source of data, and as systems grow larger and more complex, we see that the edge of networking architectures require as much consideration as any other part of the system.

Mike: When you think about what used to happen, data collection was a guy going down there on the floor about once a week with a clipboard and recording gauge readings and a stopwatch taking timings.

That would be insane today with how fast manufacturing needs to move. Today, we are able to get that data quickly, easily, and at a very affordable rate. We’re finally getting to the point where the value of that data is so much more than the cost of acquiring the data, and it just completely changes everything.

Travis: We are seeing that in the digital transformation landscape or within the fourth industrial revolution that we’re in, it’s all about data. Data is so critical to our systems, and there are a lot of surveys out there that say that 80% of data is stranded in the field and we want to unlock that.

But of course, that means that we need to plan for it. There has to be an investment that goes into it, but it’s getting more affordable and the available power that we can now put at the edge is unlike anything that we’ve had before.

What does Edge mean to you?

Cole: The edge is all around us. Anywhere you have network endpoints, connected devices; that’s all edge. For example: a plant floor, remote sites that are monitoring a pipeline, remote telco infrastructure, a lumber yard in the middle of nowhere, vehicles connected via 4G/5G. These are all examples of edge. 

The edge does introduce a new level of complexity. For example, in a typical office environment, people mostly use computers to create documents and spreadsheets, and they’re also used for video conferencing. For edge and edge-to-cloud installations, you introduce complications for the IT team that they would never experience in an office. Those can include connectivity issues and extreme environmental conditions.

A hardware conversation needs to happen because the data at the edge that is being generated (or that could be generated) is so valuable. You can start with capturing local views and reporting. It’s possible to quickly scale to business changing applications such as quality control, machine vision, predictive maintenance, or big data analysis. The starting place is getting an edge device out there to collect the data.

Mike: The world of operational technology, OT, has grown so much. Not too long ago, organizations used to ask “what of our IT capabilities can we take and extend out to the OT?” Now organizations realize the importance of OT data and want to know how to gather and leverage OT data easily and reliably. That puts a lot of challenges back on IT. 

Travis: In the “brownfield world”, that is older infrastructure, there are lots of devices that have been in place for a long time that will still be running for another 10 to 20 years. You can’t just replace these systems – that’s impossible. We have to be able to get the data that’s locked inside. We need to unleash what’s in that brownfield world into a more modern era. At the same time, you can bring in new sensors and new equipment to the edge. It’s really about both brownfield and greenfield.

When it comes to deploying Ignition, whether it’s an HMI on the plant floor at a critical asset or at a remote site, a headless data collector, a control room setup, or an on-site server at remote locations, all of that is the edge. Edge is a critical consideration for any deployment; how else are you going to get the data you need?

There are a lot of key benefits for edge to cloud deployments that stem from direct interface with on-site equipment:

• Reduced latency
• Distributed systems/scalability
• Redundancy/failover
• Reduced transmission costs
• Data loss prevention
• Local functionality

Cole: How do you scale your organization from a process standpoint and value standpoint? The applications you can deploy, from an organizational standpoint, are dramatically business changing.

Mike: For example, think about how difficult this whole thing can be at times; think about a PC acquiring data out on top of an oil head somewhere deep in the Texas desert. Keeping that connected can be a challenge. You can’t run ethernet to it or use wi-fi there. You need to use 4G and if it goes down, you’d better be able to store that data so you can analyze it later. On top of that, you need something reliable because you can’t be rolling a truck out there every week when the computer overheats.

So you’ve really got to look at that complete lifecycle, all the reliability. Can I stay connected? Can I get my data? Because if you can’t and it becomes a burden, you’ll lose the corporate support. Do it right the first time. 

Edge to cloud architecture examples

Travis: I think the applications are really crucial. Hardware is a necessity and we have to pick the right one. But the applications drive some of these solutions. And so, as we said, that could be a standalone HMI.

Slide courtesy of Inductive Automation

A lot of customers do this; they just need to connect the local equipment and provide the interface. But when we do that, we look at that as an opportunity because it doesn’t have to just be the HMI; there’s critical data that it’s collecting and working with.

Slide Courtesy of Inductive Automation

When we look at a broader solution that’s bringing data up from the edge to cloud services or a centralized data center, we can now store and forward the data. We don’t lose that data and we have that local interface. We’ve always been talking about these architectures at Inductive in terms of a hub-and-spoke model and now we’re just defining it more specifically as “edge”. 

Slide courtesy of Inductive Automation

Data collection in particular is leveraging the edge and looking at technologies like MQTT to publish that data up to a centralized system efficiently when data changes. This is perfect for remote locations where connectivity is through cellular or 4G. The technologies are in the landscape today; they are available, and along with the compute, you have everything you need to make this happen.

Slide courtesy of Inductive Automation

You put it all together: data collection, local interfaces, and then later looking at machine learning or analytics. There are some really cool real world use cases that are being done today.

Cole: I want to point out that the cost to get a POC for this type of application is very low. You can do a trial of Ignition, and get a piece of hardware from us connected and you can immediately start to see value.

Real world implementations

Let’s talk about how we’ve done that in a real world brownfield application with legacy equipment. The first thing that can be done is display and visualize the data locally and do some reporting. There are still a lot of people in the industry that walk around with a clipboard checking valves and collecting data manually. With Ignition, they can use the legacy communication protocols and take advantage of the built in drivers. Data can be pulled into Ignition from the machinery without having to replace anything. OnLogic’s industrial PCs are ideal for this type of application because they are designed for industrial environments like factories which are not usually the cleanest of environments.

I worked on a project for a chip manufacturing facility. You can imagine the factory is not an ideal environment for a computer; it’s full of chip debris, salt, chip seasonings, and more. That’s why you need a specialized piece of hardware – an industrial PC. But you can use it just like you would any computer. It’s an x86 based computer with operating systems that you know and are familiar with.

We implemented Ignition and immediately had data for plant operators giving them real-time feedback. We generated reports that provided true value for that customer.

And now we’ve created a situation where we’re showing immediate value, getting additional value of the machinery and we had a baseline. Next step for them is to scale and deploy additional edge compute resources. For example, you can see the cameras in this photo. Pair them with a higher spec device and you could do machine vision quality control or predictive maintenance. With that, you could know when it’s time to replace equipment before it fails, helping to reduce downtime. 

There are easy ways that you can start to show immediate value and prove this out with your organization. As you scale up you can look at the cloud, big data tools, dashboards, and all the other tools you can use with Ignition.

Wastewater edge to cloud example

Cole: Let’s look at how we can solve challenges at more of a distributed network. I’m going to talk about wastewater, but this could be anything from wellheads, utility grids, and more. 

In this environment, there are some additional challenges such as the temperature. In an indoor space like a factory, it’s fairly temperate. But when you get out in the field, you probably don’t have AC or heat. You need an edge device with a wider operating temperature. Other big considerations are connectivity challenges and data transmission costs. Putting the compute right at the edge enables you to send just the processed data and not the whole dataset up to the cloud or offsite server. This is really important because of the expense to transmit data. Edge computing can help to minimize that cost.

There are many applications where the PC is a critical part of the infrastructure. It’s critical to plant operations, so having something to operate outside of the network is key. As reliable as networks are, you always want to minimize the risk of downtime. This can be accomplished through Ignition Edge, where you can have the PLCs report into the computer itself and run the local HMI. If that connection to the remote server off site or cloud data center is lost, the PLCs can still transmit data and the local HMI still works for the operator. When that connection is reestablished, the tags can be sent up with no impact to plant operations.

In terms of scaling the application, whether it’s predictive maintenance, organization wide reporting, or computer vision, the long term goals and applications can work, whether it’s brownfield or greenfield, whatever the industry. But you need to consider those additional hardware and software challenges. It’s something OnLogic and Inductive can help you do. We work for customers like this all the time to solve these challenges.

Hardware and software considerations

Travis: When the edge is a critical part of the infrastructure, what does reliability mean? What is the cost to the customer in hardware, labor, and lost productivity? And how much does one hour of downtime actually cost? It’s very important to measure what downtime really means to a customer.

Cole: This is one of the most important considerations when doing an edge to cloud deployment. You need to consider the total cost of ownership for the hardware in the overall solution stack. It can sometimes be easier to say to just look at the software – if we implement this kind of application from software, we will save money from reduced downtime or improved operating efficiency. 

When looking at the total cost of ownership, the hardware can’t be ignored. If you deploy the wrong hardware it can end up costing a lot. Consider an off the shelf computer from a big box store; the computer components aren’t designed to be on 24×7. It doesn’t have the environmental protections that you need. When you look at an hour of downtime, the cost can vary. I’ve heard anywhere from $50,000 to well over $1,000,000 per hour.

There is a well known example of a large electronics manufacturer that suffered a thirty minute power outage at one of their plants. The estimated losses were over 43 million dollars. 

I was working with a company out in the southwestern United States. At the time, they were deploying a cheap PC that wasn’t meant for the environment that they needed it for and it would fail all the time. Every time it failed, a tech would pick up a new computer and drive out to the site in the middle of nowhere to fix or replace the system. It was so far, he would end up spending the night racking up even more expenses.

When they came to OnLogic, they told me their price point and I said, “let’s do some quick math.” When you buy a system that wasn’t designed for that environment, you might save a little bit of upfront cost. However when it goes down, you not only have to replace the system, but you also have to pay for two days of a technician’s time and expenses, plus the cost of downtime at that remote facility. The cost can really add up.

There’s a great case study on our website about a company called American Woodmark. They used desktop computers in their manufacturing space and the fans would get clogged up with sawdust. They had a full time employee whose job it was to open up the computers and vacuum them. When they switched to our fanless computers, they were able to reprioritize that technician to a more meaningful career and they saved money because they didn’t have to replace the computers after the vacuum damaged them. 

Hardware considerations for the edge

Travis: It reminds me a lot of advice from my parents, which is buy once; for example, buy the better pair of shoes because they’ll last longer. It’s similar here – you need to consider the long term ROI and the value of the data and not so much the initial investment. The point is to make the investment; it will protect the future.

Mike: There are 5 general considerations you should think about when selecting your hardware.

1. Environment
2. Industrial I/O
3. Regulatory requirements
4. Mounting options
5. Connectivity

Let’s start with the environment in which you install your computer.

Top environmental considerations: 

• Dirt and dust
• Temperature variability
• Power input and protection
• Impact forces and vibration
• Chemical
• Embedded lifecycle

We’ve talked already about dirt and dust like at the potato chip factory, and we’ve also touched a little on the temperature and how it’s important to consider your installation environment. Don’t forget to consider the extremes. For example, I worked in a non temperature-controlled warehouse when I was a teenager, and we used to joke that it was air conditioned in the winter and heated in the summer. The temperature was extreme and we were indoors. Industrial computers have an extended range and rugged computers can go in even more extreme environments. 

You need to also consider available power. AC power is not everywhere. For example, if you’re in a vehicle or a forklift, you will probably run DC power and it may not be the cleanest. So make sure that you’re thinking about power protections.

Cole: Looking at the most common use cases, they’re often running it in an enclosure. You have a 24 volt supply. If you’ve got terminals on that, you want a system that’s got terminals on the other side to make integration easy.

Mike: When you need a computer for railway or auto, on a forklift, you need to consider impact forces and vibration. An off the shelf computer was not designed to endure vibrations let alone the impact force of a truck hitting a pothole. Consider the worst case scenario. 

The ability to withstand chemicals might also be a consideration for you. For example, we have a customer that makes kiosks for car washes, and they are exposed to many chemicals in the air. We also have customers in manufacturing that have to spray down their equipment for cleaning with caustic chemicals. 

Finally, consider an embedded lifecycle. Once you have verified a PC, you don’t want the PC to go EOL. A lot of consumer desktops only have an 18 month lifecycle. You should consider an industrial PC with a much longer lifecycle. 

Industrial I/O

It would be a wonderful world if everything was connected to high speed Ethernet but that’s unfortunately not the case. There are many legacy systems that still have serial connections. Think about how many I/O connections you need and which variety. 

That’s the one thing that we really pride ourselves on – we build to order. We have a lot of options on the front end. If it gets a little overwhelming, just give us a call. We’ll work with you to make sure that you can meet your needs at a good price point with even room to expand. 

Regulatory requirements

The one thing I’ve seen derail more deployments than anything else is when regulatory requirements are not taken into consideration. Do the research up front and consider where you are deploying.

Cole: It’s so important to talk to us upfront about this. We have systems that are UL rated out of the box, but some countries have different requirements. At OnLogic, we have an entire team dedicated to helping customers solve these challenges. 

Mounting your industrial PC

Mike: Mounting a PC to keep it out of the way but having it still be usable and supportable is important to think about. In some environments, it would be very safe to put your PC right on the back of your monitor using a VESA mount or a wall mount. However, in industrial situations, you’re probably going to have something that looks like the DIN rail that’s pictured here. 

Think about enclosures and heat inside that enclosure; are you going to have your panel PC flush with the face of your enclosure? What does that mean? And what does it mean in terms of access? There are many questions to address when it comes to mounting. We also offer servers with more processing power available with a rackmount option.

Connectivity Considerations

It’s one thing to have the data, but you also need to have access to it. Hard-wired ethernet will always give you the best results, but there’s plenty of places that can’t be wired for ethernet. In that case, you can consider wifi or cellular. Generally, it’s best to have more than one way of getting to your data. 

Cole: One of the things that’s great about cellular is that it gives you connectivity outside of your customer’s network. So if a network is private, cellular is an easy workaround.

Factor 200 Series industrial Raspberry Pi

Mike: I wanted to show off our newest computers – the Factor Series! These industrial Raspberry Pi units are small, powerful, and give you a lot of bang for your buck. The Factor 201 is available now with Ignition Edge preloaded. Our newest model, the Factor 202, will soon be available with Ignition Edge preloaded.

Cole: These systems are a cost effective way to get started. 

Software considerations

Photo courtesy of Inductive Automation

Travis: Just as there are hardware considerations, there are also software considerations. You need software to be able to scale. You can always scale by adding more, but when it’s built in, when you think about it from the beginning, it lays that groundwork for low cost scaling. With Ignition, you can get more data, do more functions, and have more available. You need to think about the future; you need to figure out how and where you’re going to get your data and where it’s going to need to go.

Mike mentioned the connectivity in terms of cellular, 4G, and wireless; that’s obviously very important to getting the data out. However, we also need to look at how to get the data in. What are we talking to, legacy PLCs? What protocols are we dealing with? Is it OPC-UA? Are we going to be sending data through MQTT? Are we going to the cloud through cloud connectors? You need to create a roadmap to your success.

And then lastly, of course, do you have to have redundancy? How critical is the system? Do you need that redundancy or fail over? 

Edge to cloud – going forward

You want to set yourself up for future success. The investment that you are making today lays the foundation for the next five years or longer.

Ask questions like: 

• What can the edge be going forward? 
• What investment am I making now that lays a foundation in five years for a changing environment, problems, solutions, needs, so you’re not replacing the system? 

Start thinking about more types of functions and data in the future and use the tools that are provided in the cloud to supplement deployments such as cloud native tools, dashboards, and analytical functions.

Cole: You can start small, but don’t stop after that first deployment. True digital transformation is possible at an enterprise level. If your ultimate goal is to deploy a more complex application, consider that upfront and ask yourself – what does that need look like in terms of hardware and software in the short term? How about the long term? 

Mike: When we work with a customer on a POC, we don’t want to over-spec a system, but we don’t want to under-spec it either. There should be room to expand. You should be saying, what else will I be doing at the edge next year? What about 3 years? Make sure to leave yourself some room there, because the one thing that we know is that the edge is not going to get simpler. The edge is absolutely going to get more complex. Do your homework carefully and make sure that you’re thinking about the future, not just today.

Travis: We are seeing more and more need for the ability to run very complex machine learning models at the edge. These are the kinds of discussions that we’re having and we want you all to hear what those considerations should be.

It’s never too early to try it out for yourself. You can go to the Inductive Automation website and download a trial version of Ignition and try it for free. It takes about 3 minutes to download and install.

It’s simple, easy, and you can evaluate an entire system to see what Ignition can do for you without having to purchase anything, making it easy to get started. We have a great university to learn about how to do things in Ignition. There’s a comprehensive collection of videos along with the documentation system.

You can also contact either a distributor or our account executives to learn more about Ignition. OnLogic has similar ways of getting people started with your hardware. 

Cole: Our TryLogic prototyping program lets you test a system for 30 days, risk-free. If it’s not quite right, you can return it for a full refund and try another solution.

Want to learn more about the benefits of edge to cloud deployments before you start your project? Subscribe to our email communications to keep in the know on technology trends and know-how.

Editor’s Note: Inductive Automation has ended their Ignition Onboard program. Ignition licenses must now be purchased directly through Inductive Automation. While the IGN versions of our solutions are no longer available, our computers remain a great fit for use with Ignition software. Explore our recommended hardware here.