Understanding Fog Computing vs Edge Computing

By ·Categories: Tech Explained·Published On: April 15th, 2022·7.1 min read·

What is fog computing vs edge computing? You might hear these terms used interchangeably, but there is a difference. Check out our fog vs edge video below, or continue reading.


In previous blog posts, we’ve touched on edge computing and cloud computing. In this post we’re going to take a step back and look at the bigger picture to examine edge vs fog computing. We’ll explore the differences and similarities between them, and give some practical examples to try and help demystify what has become a common question as businesses of every shape and size work to establish the best location for computing power. 

So, before diving into edge computing vs fog computing, it’s important to understand cloud computing. A simplified definition of cloud computing is computing power that’s made available as an online service, frequently offered by a third party. A good example is online storage and file management providers (Google Drive is one such service). 

With these providers, your physical device does not actually store your files. The “cloud” stores your files. For industrial applications, this data, which can come in many forms, might originate at IoT sensors, and then be sent to a cloud service such as Amazon Web Services or Microsoft Azure. The physical devices in the field need to transfer the data to the cloud. This is where edge computing and fog computing come in.

What is edge computing?

Edge computing, as the name implies, brings data computation closer to the “edge” of the network where data is produced. This can help to lessen or even eliminate the need for a remote data center since all of the data is processed on-site. Latency is also reduced, allowing for real-time decision making since there’s no need to wait for the data to be processed in another location.

To better understand edge computing, let’s look at a real-life example of predictive maintenance in a manufacturing environment. With sensors embedded in the manufacturing equipment, data can be continuously sent to a nearby edge server. 

Using AI algorithms based on historical data, this data can then be processed and analyzed to determine the likelihood of a system malfunction. This, in turn, can help to prevent unplanned downtime. In fact, predictive maintenance can help to reduce unplanned downtime by as much as 70%.

You can read more about the benefits of edge computing in our blog.

Edge computing challenges

Although edge computing is ideal for many applications (especially those involving time-sensitive data), it’s important to also understand the potential downsides of edge technology before implementing it as a solution.

Up front infrastructure cost can be high with edge computing. Unlike cloud computing, which utilizes an off-site, often third party-provided “cloud” to store data (more on that below), edge computing processes and stores data locally. This means that you need more on-site hardware with edge computing.

Edge computing also carries data storage risks. If you have on-site edge devices storing information that has not been backed up, you can permanently lose your data if the hardware is damaged, destroyed, lost, or stolen. To help mitigate these risks, you should always back up your data reliably and ensure that hardware reliability is a key consideration when selecting edge devices.

What is fog computing?

Fog computing, also called fog networking, is a compute layer between the cloud and the edge. Where edge computing might send huge streams of data directly to the cloud, fog computing can receive the data from the edge layer before it reaches the cloud and then decide what is relevant and what isn’t. The relevant data gets stored in the cloud, while the irrelevant data can be deleted or analyzed at the fog layer for remote access or to inform localized learning models.

A real-life example of fog computing would be an embedded application on a production line, where a temperature sensor connected to an edge server would measure the temperature every single second. This data would then be forwarded to the cloud application for monitoring of temperature spikes. Imagine that all of the temperature measurements, every single second of a 24/7 measurement cycle, are sent to the cloud.

With a fog layer, the edge server would first send the data to the fog layer over a localized network. The fog server would receive this data and, according to certain parameters, decide whether it is worth sending on to the cloud. The end result is reduced traffic. 

For simple temperature readings, these data savings might seem negligible. But imagine if you were constantly streaming complex information or large files like images or video. The impact on bandwidth and latency could be massive depending on the application.

Edge and Fog Computing Layers - OnLogic

An example of how the sensor, edge, fog, and cloud layers of a computing infrastructure connect.

What are the benefits of fog computing?

Now that we know that fog computing is an extra layer between the edge layer and the cloud layer, what are the benefits of having that extra layer? The initial benefit is efficiency of data traffic and a reduction in latency. 

By implementing a fog layer, the data that the cloud receives for your specific embedded application is a lot less cluttered. Where a cloud would have to first weed through a pile of unnecessary data before taking any action or returning results, it can now act directly upon the data that it receives from the fog layer.

When looking at the bigger picture, there are a lot more benefits. The amount of storage you would need for your cloud application would be considerably lower. This is because the cloud would only store and process relevant data. The data transfer would also be faster because the volume of data being sent to the cloud would be significantly reduced. 

What are the disadvantages of fog computing?

One thing that should be clear is that fog computing can’t replace edge computing. However, edge computing can definitely live without fog computing. Thus, the downside is that fog computing requires an investment.

It is a more complex system that needs to be integrated with your current infrastructure. This costs money, time, and knowledge about the best solution for your infrastructure. Fog computing isn’t an ideal solution in every scenario, but the benefits can be attractive for those currently using a direct edge to cloud data architecture.

Fog computing vs edge computing: what are the key differences?

Fog computing and edge computing share a lot of similarities. Essentially, both are enablers of data traffic to the cloud. As we explained in our blog about what edge servers are, edge computing happens where data is being generated, right at “the edge” of a given application’s network. 

This means that an edge computer connects to the sensors and controllers of a given device and then sends data to the cloud. However, this traffic of data can be massive and inefficient. Irrelevant data might be sent to the cloud in addition to the useful information that’s actually needed. 

Unfortunately, even the cloud has its limits in terms of capacity, security, and efficiency when connected directly to edge devices. Enter fog computing.

Can you use the same hardware in both fog computing and edge computing?

In terms of hardware and the type of computers you can use, you can easily use edge computing hardware for the same purpose as a fog server. The difference is in where and how data is being collected and processed, not necessarily the hardware features and capabilities. 

If you take the Karbon 800 for example, which was initially designed for edge computing, it would be just as suitable for fog computing. Of course, every project is unique. It’s important to have a clear view of your overall project requirements when selecting and configuring any hardware solution. 

Fog computing vs edge computing summary

In a nutshell, edge computing is data computation that happens at the network’s edge, in close proximity to the physical location creating the data. On the other hand, fog computing acts as a mediator between the edge and the cloud for various purposes, such as data filtering. In the end, fog computing can’t replace edge computing, but edge computing can live without fog computing in many applications.

Have questions about hardware requirements for edge or fog computing? Talk to one of our specialists to find out more about OnLogic’s hardware offerings.

We originally posted this blog on June 23, 2021. We updated this blog on January 15, 2022.

Get the Latest Tech Updates

Subscribe to our newsletters to get updates from OnLogic delivered straight to your inbox. News and insights from our team of experts are just a click away. Hit the button to head to our subscription page.

Share

About the Author: Andrew Overheid

Andrew Overheid is the Marketing Technologies Manager at OnLogic. Besides making websites and creating content, he can be found at home playing the guitar. You can follow Andrew on LinkedIn.