Welcome, cheat sheet fans, to the latest chapter in our IoT-elucidating journey. Today, we tackle Edge Computing – the data-processing wonder that lives between application and cloud.
Name: Edge Computing
Also known as: Fog Computing
IBM product: IBM Edge Solution
Super power: Exploits available compute facilities on edge devices to distribute Cloud-based services – both reducing network latency and increasing data protection.
Edge Computing in ten seconds
Ready for your ‘official-sounding definition’? Here you go:
‘Edge Computing (also known in the industry as Fog Computing) is a natural and logical extension of the Cloud to exploit compute resources and data introduced on edge devices.’
How does it work?
OK. So what does that actually mean? Well, it has a lot to do with the emerging growth in computational capacity on edge devices.
Let’s begin with the idea of an edge device. We’re all familiar now with mobile devices – smartphones, tablets, even laptops. We take it for granted that these devices have significant compute, memory, storage, and network capabilities. There are approximately 3.5 billion smart phones in the world today.
What may not be as obvious are the approximately 50 billion other intelligent devices that we expect to see in the world by 2020. What started out as IoT devices, like instrumented traffic lights, sensors buried in bridges, and ovens that you can turn on from your App, are now gaining more and more intelligence. Your refrigerator can now use a camera to determine what food you are storing there, how old it is, and what you might be missing to make that casserole tonight. You can download and run Internet apps on your smart TV.
Your car has 10s and sometimes 100s of computers controlling different functions and capabilities, and these are all connected by a network. Industrial machines are being monitored a thousand times a second for any change in their performance, and then tuned in real time to optimize their output, or to predict their potential to fail – well in advance of that failure actually occurring.
This surge in computational capacity creates an opportunity for Cloud computing.
Edge Computing in action
Let’s imagine an industrial IoT application, like smart traffic lights, for example. The traffic lights capture streaming data from cameras and queue detectors in the road and send it to a central control system. The control system can understand traffic flow from this information and adjust the lights accordingly.
The data from the queue detectors has to be processed and responded to somehow, so it’s usually sent to the cloud, or to a central data warehouse. All well and good. The problem is that the cloud has to handle an extraordinary amount of data, meaning that response times may not be as swift as we’d like.
There’s also the question of distance. If the data from the queue detectors is processed and analyzed in the central control system, or in the cloud, it has a long way to travel. When there’s a large physical distance between the user or application (in this case, the lights) and the cloud, or the central control system, you get something called ‘transmission latency’. That means that the data has a long way to go, so it takes a while to travel. Consequently, response times increase and the user’s left hanging.
This is where an Edge Computing platform, like IBM Edge Analytics in Watson IoT Platform comes in. The platform allows some application processes to take place on an edge server that sits between the cloud and the user. This means that some of the workload from the cloud or user’s device can be offloaded to the edge server, close to the user, for processing. The smaller transmission distance means responses are faster. There’s a security advantage too, as it’s easier to keep sensitive data close to the network core, where firewalls and other protective elements come into play.
An Edge Computing server reduces the cost of data transmission
The advantage of Edge Computing
In a nutshell, Edge Computing can:
- Decrease the volume of movable data, which means:
- Reduced traffic
- Reduced cost of transmission
- Improve security and protect privacy, by:
- Moving encrypted data closer to the network core, closer to firewalls and other safety nets
- Cope with intermittent connectivity
- When applications run part of their computation on the Edge, they are less likely to be disrupted by limited connectivity or remote locations
- Scale easily through virtualization
Use cases for Edge Computing: predictive maintenance
Since Edge Computing has the benefit of quick response times, it’s of especial value in situations where we need to analyze real-time data from IoT devices. Manufacturing is a good example of a time-sensitive use case. Faulty equipment can lead to costly downtime and, in occasional instances, to worker injury. Defects need to be spotted and addressed quickly before they cause costly or dangerous problems.
Here, Edge Computing can help. Instrumenting factory equipment with IoT sensors and analyzing the data from those sensors at the edge, means time-sensitive issues can be dealt with quickly. The Watson IoT platform can interpret this data to provide actionable insights when and where they are most needed. For more complex, or longer-term tasks, sensor data can be sent to the cloud and analyzed alongside other data sources, such as imagery, video or text.
Go deeper: further resources
We hope that introduction answers some of your questions about Edge Computing. If you’re still not sure what it’s all about, our representatives are here to help. Speak to an IBM expert today and we’ll be glad to answer your questions.
You might also enjoy these resources:
Discover more at Think 2018
Our landmark conference, Think 2018, offers many sessions, demos and panel discussions about Edge Computing. Join us in Las Vegas from 19-22 March to go deeper into Edge Computing. Watson can even help design your personal curriculum, so you get the most from your experience.
Alternatively, here are some of the sessions on offer:
- Develop in the Cloud, Run at the Edge: Distributed Computing with Watson IoT Platform. Monday, 12:30 PM – 1:10 PM | Surf F | Session ID: 4356A
- Getting Started with Watson IoT Edge Gateways. Monday, 2:30 PM – 4:10 PM | Lab 17 | Session ID: 4966A
- Transforming Worker Safety with IoT and Edge Computing at Cenovus Energy. Thursday, 10:30 AM – 11:10 AM | Surf F | Session ID: 2431A
- Unleash the Edge. Tuesday, 1:30 PM – 2:10 PM | Surf B | Session ID: 4417A
- ADP: Securing the Enterprise Edge. Wednesday, 10:30 AM – 11:10 AM | Surf D | Session ID: 3267A
- Combine SAP and AI on the IBM Cloud to Gain a Competitive Edge. Thursday, 10:30 AM – 11:10 AM | Cloud and Data Campus Large Theater | Session ID: 8345A