In recent years, the rapid advancement of the Internet of Things (IoT) has brought increased attention to Edge Computing. Along with the diversification of data, there is a growing need for processing vast amounts of data, including images and videos captured by cameras as well as traditional environmental sensors.

The significantly-improved computational power of Edge devices has enabled the execution of analysis, algorithms, and processing of artificial intelligence (AI). Thanks to this advancement, Edge devices have opened new possibilities, achieving real-time and efficient utilization of network bandwidth by sending only essential information to the cloud.

The evolution of Edge computing

Edge computing means processing data near to where it's generated. This reduces data processing time and enhances real-time capabilities of devices. In contrast, with the traditional cloud-based approach, data is sent from devices to cloud servers, creating a delay before information is sent back to the cloud, after having been processed. However, improvements in computational capabilities of Edge devices have enabled devices to perform multiple processing tasks directly on the device itself.

Today, advancements in Edge computing have led to positive impacts across industries. First off, in the automotive industry, cameras and sensors are integrated into vehicles to process data in real time, enhancing autonomous driving capabilities. In the manufacturing industry, Edge devices are analyzing data on the production line, leading to improved manufacturing efficiency. Further, in healthcare, wearable devices are monitoring patients' health data, assisting the earlier detection of health problems.

Enhancement of computing capabilities in Edge devices

Recent technology breakthroughs have significantly attributed to the enhancement of Edge device capabilities. With increased processing power, memory, and storage capacity, Edge devices can now execute even the most advanced computing tasks. In addition, tasks such as AI model inference and image recognitions are also possible due to dedicated hardware accelerators and GPUs (Graphics Processing Units).

These improvements have become vital elements that allow Edge devices to handle diverse data types and deliver timely, quality insights to users. For instance, if you were to detect suspicious activities by analyzing video stream data from a security camera, you could use Edge devices to issue alerts on the spot, enabling swift responses.

Efficient transmission of data to the cloud

Another advantage of using Edge computing is the efficient transmission of data to the cloud. With traditional cloud methods, sending all data to the cloud would use up an excessive amount of network bandwidth and cause delays in communications. However, by performing preliminary data analysis on the Edge, and sending only essential information to the cloud, network congestion would be minimized while still processing and collecting data efficiently.

As the quantity of IoT devices continues to rise and it becomes necessary to manage data from billions of IoT devices an, the Edge approach becomes progressively more important. Edge computing allows instant response at the point of data generation, reducing data overload on servers, and fosters the development of new business models.

Envisioning the Future

Edge devices will take on an increasingly essential role across applications as time goes on. But it will require substantial technological progress at an application's edge to improve portability and maintain a smooth user experience. Therefore, it is implied that leveraging WebAssembly (Wasm) could be beneficial for achieving this goal. Wasm is a high-performance binary format that can be applied to embedded devices and be generated from programming languages like Rust and C/C++. This mechanism allows applications running on Edge devices to be independent of specific platforms and adaptable to diverse environments.

Additionally, the concept of 'swarm sensing' pertains to the collaboration among devices for sensor data acquisition, whereas 'swarm intelligence' encompasses devices cooperating with AI and advanced algorithms. Envisioning a future where such collaborative processing becomes commonplace, it becomes essential to advance the distributed processing technology on Edge devices.

Here are the several ways this can be considered in the future:

1. Expanded use cases

• Edge devices, which reduce dependency on the cloud, are assumed to be adopted in more use cases than ever. This will lead to a decrease in latency and enhanced privacy.

2. Enhancing portability

• Through the utilization of technologies like WebAssembly, Edge devices will boost the mobility of Edge applications, creating seamless functionality across devices and platforms.

3. Cross-device collaborative processing

• By incorporating the concepts of 'swarm sensing' and 'swarm intelligence,' devices can collaborate in real-time to facilitate data sharing and advanced processing. This gives rise to new applications based on collective knowledge and cooperation.

4. Reduction in cloud dependency

• With further advancements in Edge devices, the dependence on the cloud can decrease. This would result in decreased network traffic, facilitating cost savings, and efficient utilization of resources. Additionally, by deploying various Edge devices, it becomes possible to create a fully, self-sustained distributed environment, including applications, and potentially achieve a cloudless systems.

5. The function and significance of open-source software (OSS)

• OSS has a significant role in driving the advancement of edge computing. The community's cooperation drives improvements in new features and security and fosters the growth of the ecosystem.

To work towards this possibility, it is essential for the technology community and the industry to cooperate and create frameworks that facilitate the progression of edge computing. We should also view slight changes to cloud as a facet of technological trends, as it could potentially contribute to the evolution of Kubernetes (Kube) and WebAssembly (Wasm).

To realize this future vision, developers should cooperate and actively participate in developing and improving open-source software. We believe that the evolution of distributed processing technology through Edge computing offers an enticing goal for the future of technology, with the potential to transform our digital world. We understand that there are challenges we should actively address to collectively build this future, and we at Sony Semiconductor Solutions Corporation ("SSS"), are actively addressing them.

In order to increase businesses agility and simplicity to leverage Edge AI, it's important to have unified tools and an environment that facilitates software and application development and system implementation from cloud to edge.

Leveraging technological leadership of image sensor, AITRIOS(TM) by SSS offers an edge AI sensing platform that enables efficient development and deployment of edge sensing solutions. It can also enable you to optimize and manage your solutions with published SDKs alongside a community and greater ecosystem.

++

To learn more about edge AI innovation focusing on image sensing, join us at KubeCon + CloudNativeCon North America 2023, which will take place from Nov 6-9.

##

ABOUT THE AUTHOR

Munehiro Shimomura, Open Source Program Manager, Sony Semiconductor Solutions Corporation

Munehiro Shimomura is a senior software engineer and a seasoned leader, concurrently serving as an engineering manager. With expertise in embedded systems, he have actively participated in the development of vast range of products including televisions, video equipment, and cameras as well as standards and operating systems. Going beyond the scope of embedded development, his expertise extends to providing guidance and mentorship in specialized areas such as networking, system architecture, and security. Furthermore, his strong interests lie in open-source architecture, cloud computing, and adhering to best practices.