Edge Computing: Solving the 0.1 Second 'Deadly Delay'

Data is the “lifeblood” of modern society. Although fiber optics can transmit data at 200,000 kilometers per second (two-thirds of the speed of light in a vacuum), delays are inevitable in long-distance transmission.

For example, a signal sent from Beijing takes at least 0.1 seconds to reach Chile, which is about 20,000 kilometers away. Although this delay is almost negligible, in an upcoming world of interconnected devices, even the slightest data delay can have a significant impact on certain industries, such as remote surgery, stock trading, and autonomous driving.

Therefore, reducing the distance of data transmission has become an urgent need. This demand has also driven data storage and computation to shift towards intelligent device terminals, leading to the emergence of edge computing.

How can we understand edge computing? What are its advantages? What promising applications can we expect? Let’s enjoy the exploration:

＋

How to Understand Edge Computing Through the Nervous System?

Conceptually, edge computing is a distributed computing architecture that moves the computation of applications, data, and services from centralized network nodes to edge nodes in the network logic. Since edge nodes (such as smart devices, mobile phones, gateways, etc.) are closer to users or data sources, the speed of data transmission and processing can be effectively improved, reducing delays.

To better understand edge computing, let’s reference the human central nervous system.

Imagine that when our hand touches a flame, our first reaction is to quickly retract it, and only then do we feel the burning pain. In this process, the skin receptors first receive external stimuli and generate nerve impulses that travel to the spinal cord. Subsequently, the spinal cord simultaneously transmits instructions to the muscles in the arm to retract immediately while also sending signals to the brain cortex, allowing us to perceive pain.

Edge computing is like the spinal cord that can instantaneously make reflex actions. Due to the short transmission path, it often features fast response speed and low latency but struggles to handle overly complex information; whereas cloud computing is akin to the brain, capable of processing more complex information but often relying on longer transmission paths.

＋

The Era of Data Explosion

Four Major Advantages of Edge Computing

△Image: Edge Computing Infrastructure (wiki)

From a technical or commercial evolution perspective, the exponential growth of IoT devices has generated a vast amount of data that needs to be processed in data centers. In traditional cloud computing models, the process of data transmission and storage faces issues such as high latency, congestion during peak periods, and low bandwidth. As a complement to cloud computing, edge computing focuses on solving the efficiency issues of data transmission, processing, and storage. Therefore, edge computing is more inclined to be a new solution that extends cloud computing towards the terminal and user side.

Specifically, edge computing has the following four advantages:

High Speed

Edge computing devices in IoT can process data locally or at nearby edge data centers. Since the information collected does not need to be transmitted to traditional cloud facilities, it can significantly enhance the response speed of smart devices. For instance, in the field of facial recognition, the response time of edge computing is reduced from 900ms to 169ms, even faster than the human response time for recognizing faces (370-620ms).

Security

Edge computing decentralizes data processing, storage, and applications across a wide range of devices and data centers, making it difficult for a single attack to compromise the entire network. On the other hand, traditional cloud computing transmits private data collected from wearable devices, medical devices, and industrial manufacturing to data centers over long paths, which can easily lead to data loss or information leakage. By storing and processing data at the edge, this risk can be effectively mitigated. Additionally, the ownership of the collected data will shift from service providers to end users.

Low Bandwidth Requirements

Edge computing reduces the amount of data transmitted to cloud centers through local processing, thereby lowering the demand for network bandwidth. This feature is particularly important in bandwidth-constrained scenarios, such as poor signal conditions on airplanes or in mountainous areas.

Scalability

Edge computing provides a cost-effective scalability path, allowing companies to expand their computing capabilities through a combination of IoT devices and edge data centers.

＋

Sky and Sea

Edge Computing Opens Up More Application Scenarios

Based on the above characteristics, in conjunction with technologies like 5G, AI, and cloud computing, edge computing has a broad range of application prospects across various industries:

In smart cities, edge computing supports the computation and services of large-scale infrastructure, enabling low-latency, low-cost applications for terminal devices. For example, by equipping each subway train with a passenger counting system (PCN) and an IoT edge gateway, the camera in the counter captures the movement trajectory of passengers and transmits the “number of people” data to the gateway; the gateway aggregates data and provides GPS location data, which, once standardized, can be sent to the IoT platform, matching subway operation frequency with passenger flow, thereby enhancing operational efficiency.

In the field of intelligent manufacturing, edge computing, combined with AI, uses local sensors to control and manage output, significantly improving efficiency and reducing errors. For example, Lenovo’s Morning Star robot, supported by powerful edge computing capabilities and intelligence, enables workers to precisely perform remote painting tasks through the robot. After one natural demonstration operation, the painting capability for that component is stored at the edge, allowing the robot to autonomously paint similar components thereafter.

In satellite communication, edge computing can adapt to the environment to provide low-latency, low-bandwidth, and low-cost services. Given that many terminal devices in satellite communication are often distributed in remote environments, with long transmission distances, limited bandwidth, and high costs for inter-satellite communication, edge computing effectively addresses this issue by localizing data processing, with broad application prospects in areas such as shipping, aviation, oil drilling, mining operations, and military infrastructure.

＋

Broad Market

Rapid Development of Edge Computing in China

With the development of 5G and AI, it is foreseeable that the digital transformation of various industries, including healthcare, autonomous driving, and smart retail, will gradually accelerate, giving rise to enormous demand for IoT and further releasing the development space for edge computing.

According to the International Data Corporation (IDC) report on the Chinese edge computing server market (second half of 2020), the overall market size for edge computing servers in China reached $1.542 billion in the second half of 2020, with a total of $2.655 billion for the entire year, representing a 16.3% increase compared to 2019. IDC predicts that from 2019 to 2024, the compound annual growth rate of China’s edge computing server market will reach 22.0%, surpassing the global average growth rate of 19.6%.

Currently, China’s edge computing development has several major driving forces, with Lenovo Group being one of them. Based on its technical experience in the manufacturing field and intelligent transformation, Lenovo Group has formed a smart edge computing layout that includes edge hardware, edge infrastructure, and edge intelligence.

As an important component of Lenovo’s industrial IoT platform suite, Lenovo Edge Computing is responsible for the intelligent connection and data access of industrial or general industrial devices. It provides data collection and standardized processing capabilities based on industrial protocol conversion, as well as on-site data analysis, intelligent computing, and reverse control capabilities, characterized by safety, stability, high efficiency, and low power consumption, and can be widely applied in scenarios such as the connection and reverse control of manufacturing production line equipment, remote access and management of energy or public utility devices, and data access and edge computing in production environments. The aforementioned Lenovo Morning Star robot is a typical application in the field of intelligent manufacturing.

With its existing strength, Lenovo Group was included in CB Insights’ 2020 list of “China’s Edge Computing Drivers.” Currently, Lenovo Group continues to explore the integration of edge computing with cloud computing, 5G, AI, and more in various scenarios. Recently, Lenovo Group and China Unicom jointly released a vehicle networking solution, using MEC (Edge Computing) + 5G to improve data calculation efficiency and data integration.

As application scenarios continue to enrich, edge computing will also continue to evolve and improve. Industry insiders predict that future edge computing will require higher computing power, finer computational granularity, more complex types of computing power, and higher security requirements. Driven by both technology and application, we can expect not only faster and smoother live broadcasts and gaming experiences in our daily lives but also witness the further realization of cutting-edge technologies such as autonomous driving and remote healthcare.

＋

References

What is Edge Computing? Understand It All at Once
Huaxin Consulting: “Approaching Edge Computing”
Lenovo Group Senior Vice President Rui Yong: Intelligent Edge Computing, Bringing AI to Your “Side”
Economic Observer: Edge Computing, New Opportunity in the 5G Era
The Edge Computing Infrastructure

＋

Previous Recommendations

THE END

Related posts

Leave a Comment Cancel reply