Edge-cloud Collaboration: Enhancing Computational Efficiency

Over the past decade, cloud computing has become the cornerstone of enterprise IT architecture, driving profound changes in digital transformation. By providing centralised data storage, computing resources and services, cloud computing enables enterprises to deploy and manage their applications at lower costs and with greater flexibility. However, with the rise of the Internet of Things (IoT), 5G networks and other emerging technologies, the exponential growth of data generation and processing requirements makes it difficult for centralised processing architectures that rely solely on cloud computing to meet performance requirements in specific scenarios, especially in terms of low latency, high bandwidth and real-time data processing.

Edge computing has thus become a vital supplementary solution. By performing computing and processing at a location close to the data source, edge computing not only reduces the latency of data transmission, but also enhances the real-time response capability of the system while reducing the pressure on network bandwidth.

The wide-area coverage of cloud computing combined with the localised processing capabilities of edge computing can build a more efficient, flexible and reliable distributed computing architecture. This collaborative model is leading the innovation of the next-generation network architecture, providing new opportunities and challenges for enterprises and service providers.

In this article, we will discuss the following aspects. The navigation bar will help you quickly locate your reading needs.

Table of Contents

What Are the Limitations of the Development of Cloud Computing?

What Are the Origins and Limitations of Edge Computing?

Why Do We Need Cloud-edge Collaboration?

What Are the Collaboration Modes of Cloud-edge Collaboration?

What Are the Applications of Cloud-edge Collaboration?

Summary

What Are the Limitations of the Development of Cloud Computing?

Cloud computing is a type of distributed computing, which means that huge data computing programmes are decomposed into countless small programmes through the network "cloud", and then these small programmes are processed and analysed through a system composed of multiple servers to obtain results and return them to users.

In recent years, cloud computing has enabled big data processing. Users only need to upload data to the cloud, and use the super-powerful and efficient computing platform of the cloud computing centre to centrally process computing needs. At present, smart IoT devices are usually operated by sending data to the cloud through the network, and the cloud performs unified processing.

However, with the diversification of application scenarios and the continuous growth of data volume, the centralised architecture of cloud computing has gradually exposed some limitations:

Delay and Real-time Problems: Cloud computing relies on centralised data centres, and all data processing and computing tasks usually need to be completed on remote servers. Although this architecture has advantages in computing power, for application scenarios that require real-time response, such as autonomous driving, industrial control, smart city monitoring, etc., data transmission delays may lead to unacceptable response times and affect the overall performance of the system.

Bandwidth Consumption and Cost: With the popularisation of IoT devices and high-definition video, the amount of data generated has surged, and transmitting massive data to the cloud for processing will greatly increase the pressure on network bandwidth. This not only affects the efficiency of data transmission, but may also lead to higher bandwidth costs, especially in large-scale deployments.

Data Privacy and Security: Centralised data storage and processing expose cloud computing to greater data leakage and security risks. Although cloud service providers have taken a variety of security measures, in some scenarios with extremely high requirements for data privacy and sensitivity, centralised cloud computing architecture still has potential risks and uncertainties.

What Are the Origins and Limitations of Edge Computing?

Edge computing is a new computing model that performs computing at the edge of the network, where the edge refers to any computing and network resources between the data generation point and various paths in the cloud. The edge data is connected to the cloud service on one side and the IoT service on the other side.

With the development of IoT technology, the centralised cloud computing architecture cannot solve all IoT problems such as resource utilisation, data fusion, application compatibility, and unified operation and maintenance, which puts huge pressure on network bandwidth, computing power, and storage capacity.

The advantages of the edge computing model are highlighted:

• 1. Processing a large amount of temporary data at the edge of the network, no longer uploading everything to the cloud, which greatly reduces the pressure on network bandwidth and data centre power consumption;

• 2. Data processing is done near the data producer, and there is no need to request a response from the cloud computing centre through the network, which greatly reduces system latency and enhances service response capabilities;

• 3. Edge computing no longer uploads user privacy data, but stores it on network edge devices, reducing the risk of network data leakage and protecting user data security and privacy.

Edge computing provides lower latency and more efficient data processing capabilities, but its development is not without challenges. Compared with cloud computing, the limitations of edge computing are mainly reflected in the following aspects:

Limited Computing Power and Resources: Edge devices are usually deployed in resource-limited environments, and their computing power and storage resources are far less than those of large data centres. This resource-limited environment makes it difficult for edge computing to handle complex large-scale data analysis and computing tasks, especially in scenarios that require high-performance computing.

Management Complexity: Edge computing involves the management of a large number of distributed nodes, and the complexity of managing and maintaining these nodes increases significantly, especially in large-scale deployments. This includes not only hardware maintenance, but also software updates, configuration management, and security policy enforcement, which increases the difficulty and cost of operation and maintenance.

Data Consistency and Integration Challenges: In an edge computing environment, since data processing is distributed on multiple nodes, how to ensure data consistency, integrity, and effective integration with cloud data is a very challenging task. If handled improperly, it may lead to data inconsistency problems, affecting the reliability and accuracy of the overall system.

Why Do We Need Cloud-edge Collaboration?

With the deepening of the trend of the Internet of Things and the continuous development of network technologies such as 5G, the number of consumer Internet of Things and industrial Internet of Things devices will continue to grow. At present, most smart Internet of Things devices rely on uploading data to the cloud for centralised processing by the cloud. However, the widespread access of smart terminals and the transmission of massive sensory data occupy huge network bandwidth, and directly transmitting data to the cloud also increases the risk of privacy leakage. Computing on the edge side can reduce the traffic of the core network, thereby releasing the pressure of network bandwidth, and can also achieve certain data protection, but its limited resources will lead to the inability to meet the model accuracy requirements.

Cloud-edge collaborative computing inherits the advantages of cloud computing and edge computing to simultaneously achieve high-precision, low-consumption, fast response, and low-latency application scenario requirements.

• Cloud computing is responsible for computing tasks that edge nodes are difficult to handle. At the same time, through big data analysis, it is responsible for the processing of non-real-time and long-cycle data, optimises the output business rules or models, and delegates them to the edge side, so that edge computing can better meet local needs and complete the full life cycle management of applications.

• Edge computing is mainly responsible for real-time, short-cycle data processing tasks and real-time processing and execution of local services, providing high-value data for the cloud.

In more scenarios, cloud computing and edge computing form a complementary and collaborative relationship. Edge computing needs to work closely with cloud computing to better meet the needs of various application scenarios.

Cloud-edge collaboration is a distributed open platform that integrates communications, computing power, data storage, and application services. Compared with the global, long-cycle, high-latency, and big data computing characteristics of the cloud side, the short-cycle characteristics of edge computing can better support local services. Therefore, the edge side and the cloud side are not a simple replacement relationship, but a complementary and collaborative cooperative relationship. By building a unified and efficient collaborative framework for collaborative fields such as resource collaboration, data collaboration, application collaboration, and service collaboration, cloud-edge complementarity and resource integration can be achieved.

What Are the Collaboration Modes of Cloud-edge Collaboration?

As an integrated computing architecture, cloud-edge collaboration combines the advantages of cloud computing and edge computing in various ways to meet the needs of different application scenarios. The following are several main cloud-edge collaboration modes:

Task-sharing Collaboration: In task-sharing collaboration, the cloud and edge make reasonable division of labour according to the characteristics of the task.

• Tasks with strong real-time, large data volume and delay sensitivity are usually processed by edge computing, such as image recognition in video surveillance and equipment monitoring in industrial control. These tasks need to be processed immediately to ensure the efficient operation of the system.

• The cloud is responsible for processing tasks with high complexity, large computational volume and no need for real-time response, such as large-scale data analysis, model training and long-term data storage. The cloud can also send optimised models or strategies to edge devices to improve the processing efficiency and accuracy of edge computing.

Data Hierarchical Collaboration: In the data hierarchical collaboration mode, data is processed hierarchically according to its importance and processing requirements.

• Sensitive data and data that require rapid response are processed at the edge to reduce delays and security risks that may be caused during transmission.

• Edge computing devices can pre-process, screen and filter data first, and transmit the processed high-value data to the cloud for further analysis and storage.

Resource-sharing Collaboration: In resource-sharing collaboration, the cloud and edge maximise the utilisation of computing and storage resources through resource sharing and collaborative utilisation.

• Edge devices can dynamically request support from cloud resources under high load conditions, thereby improving their own processing capabilities and ability to cope with sudden demands.

• Conversely, when the edge device load is low, some computing tasks can also be sent back to the cloud to optimise the overall resource utilisation efficiency of the system.

Intelligent Scheduling Collaboration: Intelligent scheduling collaboration achieves dynamic collaboration between the cloud and the edge through AI algorithms and intelligent scheduling mechanisms.

• Based on the real-time network status, computing load and task requirements, the system can automatically decide whether the task is processed by the edge or taken over by the cloud.

• In this mode, the scheduling system not only considers the optimal configuration of computing resources, but also adjusts the data transmission strategy according to the network conditions to minimise delays and improve processing efficiency. This collaboration method is suitable for complex application scenarios that need to respond flexibly to changes, such as intelligent transportation systems and smart city management.

Security-enhanced Collaboration: In the security-enhanced collaboration mode, the edge and cloud jointly undertake the task of data security and privacy protection.

• Edge computing can encrypt and process sensitive data locally to prevent unprocessed sensitive information from being uploaded to the cloud, reducing the risk of privacy leakage.

• The cloud provides global security protection through higher-level security measures such as identity authentication, data encryption, and access control.

In this collaborative mode, the cloud and edge computing work together to build a more secure and reliable computing environment.

In summary, the various ways of cloud-edge collaboration provide flexible and efficient solutions for different application scenarios. By choosing a reasonable collaborative mode, enterprises can better utilise the respective advantages of cloud computing and edge computing to meet changing business needs.

What Are the Applications of Cloud-edge Collaboration?

Cloud-edge collaboration has demonstrated its important value in application scenarios in multiple industries:

• Intelligent Manufacturing: edge computing processes production line data in real time, cloud computing performs large-scale data analysis, optimises production processes and equipment maintenance.

• Intelligent Transportation: edge nodes manage traffic signals in real time, and the cloud is responsible for traffic big data analysis to improve the efficiency of urban traffic management.

• Smart Healthcare: edge computing monitors patient health data, cloud computing stores and analyses long-term data, and supports precision medicine.

• Video Surveillance and Security: edge devices process video data in real time, and the cloud stores key data to enhance security response speed and strategy optimisation.

• Smart Retail: edge devices analyse customer behaviour in real time, cloud computing supports sales trend analysis and supply chain optimisation, and improves customer experience and operational efficiency.

• Unmanned Driving: edge computing ensures real-time vehicle decision-making, and cloud computing provides global data support to improve the safety and performance of unmanned driving.

The Internet of Vehicles also has the characteristics of high-speed node movement and dynamic changes in topology. With the help of cloud-edge collaboration, the Internet of Vehicles can select computing tasks based on actual environmental conditions and limitations, complete most of the calculations in the constructed cloud-edge collaborative network, and send the results to the requesting vehicle in real time through transmission means such as roadside units.

FS PicOS® for Automated Driving Solution

FS provides a range of tailored hardware, easy-to-deploy application and management software, and end-to-end services for HPC network solutions. With this, businesses can respond to customers instantly, run networks with maximum efficiency and safety, and bring innovation in autonomous driving.

Explore the Benefits of FS HPC Network

• Standardised Operating System: The switch runs the PicOS® operating system, which accelerates feature development, enhances high availability, and ensures consistent deployment of features across the data centre network.

• Unified Management Platform: AmpCon™ enables unified configuration, monitoring and maintenance of data centre networks, thereby eliminating costly downtime and time-consuming manual tasks.

• Cost-effective Solution: FS offers original products to ensure quality. With a global supply chain, FS delivers solutions that cut costs through its rich product ecosystem.

• Solution Design & Implementation: FS solution architects and service delivery teams provide with expertise and insight, help customers quickly complete analysis and evaluation, planning and design, installation and deployment.

• On-site Service: FS offers fast and reliable on-demand field support services in the US, Europe, and Singapore, including on-site surveys, installation, and troubleshooting, etc., helping customers reduce costs.

• Same-day Shipping: With over 50,000㎡ of global warehouse spaces equipped with automated management systems, FS guarantees 90% of orders for same-day shipping and supports self-pickup.

In the FS autonomous driving solution, the working mechanism of cloud-edge collaboration is perfectly reflected:

Edge Computing

• In data centre A, the core switch N8560-32C is connected to different servers (GPU servers, X86 servers, storage servers, NAS), which are used for edge computing.

• Edge computing processes critical real-time data locally (data centre A), such as sensor data generated by autonomous vehicles, to ensure fast response and reduce the latency of data transmission back to the cloud.

Cloud Computing

• The core switch N8550-32C in data centre B is connected to X86 servers, which can be responsible for more complex computing tasks such as big data analysis, model training, etc.

• The cloud computing centre can collect high-value data sent back from edge nodes for further processing and analysis.

Cooperation Method

• Task Sharing: Edge computing nodes in data centre A process real-time data related to autonomous driving, while data centre B, as the cloud, processes large-scale data storage and non-real-time computing tasks.

• Data Classification: Edge computing nodes process and filter key data, and then transmit it to the cloud through 10G fibre links for further analysis, which reduces bandwidth pressure and improves overall system efficiency.

• Resource Sharing: Peer-Link between two data centres provides the ability to share resources, ensuring efficient resource utilisation and task scheduling. At the same time, it ensures that when one party fails, the other party can take over key tasks.

• Security Enhancement: Edge computing can complete preliminary data protection locally, reduce the risk of privacy leakage, and then transmit data to a secure cloud for further storage and processing.

  

For more deployment cases of autonomous driving networks, please click FS Helps an Autonomous Vehicle Startup to Build a Data Centre Network.

If you are planning or building an autonomous driving solution, FS, as a leading global network solution provider, can provide you with comprehensive support. Our professional team will customize a high-performance, low-latency network architecture according to your needs to ensure that your autonomous driving system achieves optimal performance in data processing, storage, and transmission. Consult us now to start the future of autonomous driving!

Summary

From autonomous driving to smart cities, the application of cloud-edge collaboration has demonstrated its huge potential and practical value. Especially in the field of autonomous driving, our solution fully utilises the complementary advantages of cloud computing and edge computing to achieve real-time data processing and analysis, ensuring higher driving safety and efficiency. With the continuous development of technology, cloud-edge collaboration will bring innovation to more fields and lead the future of the digital era.