1. Foreword

Currently, we are in the era of Cloud Computing where data will be processed in Data Centers, not on devices.

However, in the era of Internet of Things (IoT), devices such as TVs, refrigerators, light bulbs, etc. are all connected to the internet. With a huge amount of data from billions of IoT devices like that, data transfer to Cloud for processing and returning results can cause delays (it may be only one per thousand seconds) or data security are issues being considered. To solve this problem, Edge Computing was launched.

Edge Computing is a distributed computing model, bringing computational processing and data storage closer to the necessary location to improve speed and save bandwidth (according to Wikipedia’s definition).

This is an inevitable trend in the IoT 4.0 technology era, which is forecasted by Gartner as one of ten technology trends that will impact and transform economic sectors from now to 2023.

Figure 1. Top 10 technology trends forecast by Gartner in 2019.

2. Edge Computing Architecture

According to Cloud Computing model, Edge Computing is located in the middle layer close to IoT devices in the classification model below.

Figure 2. Edge Computing Architecture.

From the classification model above, we can see that:

  • The top layer is the cloud data centers to process and analyze complex tasks such as Big Data, Machine Learning, etc.
  • The middle layer is Edge Computing: It can be seen that the computational layer is located next to or near IoT devices to connect and process local data of billions of IoT devices. The term “Edge Computing” is used to describe computing centers between the Cloud but near the devices, called the Edge. In addition, it refers to the compution border between the internet environment and the local network environment. Some documents or articles that use the term “Boundary Computing” or “Margin Computing” may not sound better or more difficult to understand than the term “Edge Computing”.
  • The final layer is IoT devices such as sensors, controllers, etc.

3. Benefits of Edge Computing

Figure 3. Main benefits of Edge Computing.

As mentioned in the introduction, IoT technology era has billions of devices that will be connected to the internet, local data processing on the spot, security, transmission latency, etc are the advantages that Edge Computing brings. Let us analyze and evaluate the five main advantages that Edge Computing brings as follows.

Speed:

The first benefit of Edge Computing is to increase network performance by speed. Because Edge devices are located next to IoT sensors or in nearby data centers. Collected information is not transmitted to Data Centers far from thousands of kilometers with hundreds of network devices in the middle (switch, router, etc.). Furthermore, with existing fiber optic technology, it allows fast data transmission by two-thirds of the speed of light.

By collecting and processing such local data and reducing physical transmission distance, Edge Computing can significantly reduce latency. The final result is the higher speed with latency measured in microseconds instead of milliseconds.

Security:

With Cloud Computing, data must be transmitted to data centers for processing. This can cause certain security vulnerabilities, enabling hackers to capture packets on the route. Although most communication methods are currently encrypted, there will still be certain vulnerabilities and weaknesses, as long as hackers catch a piece of the packet, they will also find the way to hack the entire system.

In contrast, with Edge Computing, sensitive, important data will be processed right at the internal device without being sent, thereby contributing to better protection of your data.

Reliable:

It is not surprising that Edge Computing provides better reliability, because Edge devices are located next to IoT devices or located in nearby data centers capable of storing and processing data locally, ensure components continue to work normally and data is not lost even if internet connection is lost.

Figure 4. Edge Computing ensures reliability and data even when Internet connection limitation.

Cost-effective:

With a huge amount of data from billions of IoT devices as mentioned, transferring this entire data to Data Centers will consume considerable bandwidth, meaning that the transmission cost will be high. However, the application of Edge Computing, the transmission of the entire data is not necessary and allows the enterprises to decide which services or data will be processed and stored locally, which data will be sent to Cloud.

Scalability:

Edge Computing allows for easy scalability by adding Edge devices whenever they need to connect IoT devices increases but does not increase bandwidth significantly.

4. Some examples of Edge Computing applications

– Intelligent traffic camera system

Figure 5. Application of Edge Computing in Intelligent traffic.

Imagine, in the big cities there are hundreds of intersections. Each intersection is fitted with 3 to 5 cameras to monitor traffic. With hundreds of such intersections, the number of cameras installed is huge. If transferring the entire video to the Cloud server for processing, it may cause congestion, and the cost of renting the bandwidth will be very expensive.

Moreover, video signals from these traffic cameras may be covered by state secrets (for example in Vietnam), so it is not possible to transmit this video to the Supplier’s Data Center. The application of Edge Computing technology can install Edge equipment directly connected at the camera or build a separate data center under the control and supervision of the Police, to process and store data immediately in place. Only metadata is transmitted to the Cloud server for analysis and processing and then sends commands to control the traffic lights.

– Edge Computing with AWS IoT for Connected Vehicles Solution

Figure 6. Edge Computing with AWS IoT (Amazon) for Connected Vehicles Solution.

Understanding this trend, the “Big-Four” group on Cloud Computing services (including Amazon, Microsoft, Google, and IBM) is planning to gradually expand their data centers around the world in order to be more decentralized and closer to the device, regardless of location.

For example, Amazon provides the AWS IoT platform for Connected Vehicles Solution. With AWS IoT platform, it allows car manufacturers or developers to build IoT applications on data handling in cars such as detecting and sending abnormal warnings to vehicle owners; monitor history, itinerary and send periodic maintenance notice; diagnose vehicle errors; monitor and send information about states such as fuel, tire pressure, battery voltage, etc. All of the above information collection, analysis, processing and manipulation are done right on a car-based device, without having to manage any IT infrastructures.

Data on cars are authenticated and encrypted on connected devices before they are transferred to Cloud.

The above are just a few typical examples, besides that Edge Computing is considered a breakthrough technology, can become a “world-changing” tool in the field of IoT, along with the data rate of the network 5G, they will have an important role in areas such as smart traffic, smart factories, smart health, smart home, etc.

5. Conclusion

With the rapid growth of IoT devices introduced every year, the launching of Edge Computing will bring about leaps of development in speeding up data processing and transport, as well as security. But Edge computing is not a substitute for Cloud Computing, because Cloud Computing has its own benefits, they are complementary architectures and come together to create powerful platforms for the IoT industry. This will impact and promote the transition of economic sectors from now to 2023.

MSc. Hoang Van Cuong (Mr.)
Solution Research & Development Specialist (ISS)
FPT Information System Company

Related posts: