Unraveling the mysteries of edge computing
The world is digitizing and there are a lot of new technologies out there that can make our lives more simple. We all know this. However, most of us don’t have the faintest idea how these technologies actually work. At Edgise, we specialize in edge computing, which is implementing AI on the edge. Naturally, we often get the question: “What is it you actually do?”
In this article, we will try to answer that question by giving you an easy explanation on edge computing. Let’s dive in.
AI, in the cloud or on the ground?
Edge computing, defined
According to Gartner, edge computing:
… is part of a distributed computing topology where information processing is located close to the edge, where things and people produce or consume that information.
Let’s unravel this definition, shall we?
Close to the edge
In the past, companies mainly used artificial intelligence (AI) inside the cloud, as this was the only way to power these kinds of algorithms. The way this works, is that raw data, for instance from drones, smartphones, or cars, is sent toward the cloud. Then, the AI model analyzes the data and sends the analysis back to the local devices.
As the computational power of most devices has increased significantly, it became possible to implement AI models on the devices instead of uploading it into the cloud. And so, edge AI was born. This means that the AI algorithms perform their analyses locally. Data collection and data processing both take place on the same hardware. This means that the hardware does not need to be connected to your network.
The future of edge computing
In the coming years, edge computing will play a crucial role in moving data away from the cloud. As rapidly increasing amounts of to-be-processed data are driving network capabilities to the impossible, processing them on the edge offers a solution.
Why? Edge computing only sends essential results, without the need to transfer all of the raw data. In other words, this kind of technology makes sure that only the relevant data is processed within the network. This ultimately reduces latency issues. By moving artificial intelligence from the cloud to the edge, we shorten response times, increase performance, and reduce privacy risks.
Edge computing in the field
Counting algorithm
We have been developing a general counting algorithm based on AI for some time now. In essence, the algorithm makes it possible to count and classify people, vehicles, or other objects in real-time, based on camera images.
We designed the algorithm in such a way that it runs on small electronic devices located close to the camera. As a result, we can actually process the images locally on the device and we do not have to keep sending them to the cloud or another type of system. On the one hand, this eases the pressure on networks, makes it possible to count in places where no strong internet connections are available and prevents privacy problems if people come into the picture.
On the other hand, it ensures a networking cost reduction, because there is no specific need of a network. Another advantage is that we don’t need to travel to the cloud and back, so in general the application runs much faster. This is important when you want to run your application in real time, for example. In addition, edge computing also provides additional protection against hackers. As the data does not go up into the cloud, hackers have less chance to hack the data. In turn, this results in more privacy. ✔️
Smart cities
In the case of smart cities, there has been an increasing interest in intelligent algorithms. These can monitor objects and situations based on raw data of cameras, microphones, and all types of sensors. However, the question of preserving privacy, lowering network costs, and keeping existing sensor infrastructure seems to be tough to answer. Luckily, edge computing can offer a solution to this problem, as it keeps privacy, latency, and infrastructure in mind.
Virtual store agent
We also saw the opportunity to use edge computing and edge AI to keep people safe. During the corona crisis, we developed our counting device Telly. This device can be used to count the capacity in, for example, a store. Telly can be set so that when there are too many people in a room, an alarm bell rings.
Telly
In 2020, we implemented Telly at Carrefour. We installed Telly at the entrance of the store, so a real-time AI algorithm would analyze people coming in and out of the store. If the store was too crowded, the doors would close automatically and Telly showed an informative message on the kiosk next to the entrance. If people forgot to wear face masks, the doors would stay closed and the system would inform the person to put on a face mask before entering the store. In essence: a helpful solution in these times. Interested in how we did this exactly? Read the full case study here.
As you can see, there are many possibilities with edge computing. Want to see how the technology can help your business? Just schedule a call with us here.
