Analog/Mixed Signal

What are the major applications of edge computing?

12 October 2022
Some of the potential edge computing applications. Source: Yingyaipumi/Adobe Stock

Some modern computing and design trends tend to be overhyped before they find their place in an evolving technological landscape. Today’s sometimes-hyped, sometimes-misunderstood computing paradigm is edge computing, where processing power for end-user applications is offloaded from the cloud and brought to the edge of the network. It’s a simple concept that enables better service delivery in many applications that require low latency, and it is bringing about a new model for compute access in embedded devices.

The applications outlined below provide some guidance and examples where edge processing can be used as an enabler of better service delivery for end users. Ultimately, the goal is to provide client devices access to high compute capabilities without increasing the complexity of the client device and without adding to network congestion.

The edge computing model

The edge computing model is a critical enabler of time-sensitive applications that also require high compute. When high compute density is not available on an end-user device, it might make sense to implement those tasks at the edge. The applications listed below are time-critical while requiring high compute, but with end-user devices that may not have the compute density required to perform important processing tasks. These are some areas where edge computing can offer a solution.

Time-critical application examples

Sensor fusion — New IoT systems, as well as many systems in automotive and aerospace, are incorporating more sensor interfaces that require significant processing power to make sense of the wider world. In sensor fusion, edge computing systems can provide that processing power directly by aggregating data from multiple end-user devices. As an example, pre-processing on an end-user device allows data size reduction so that it can be quickly transmitted to an edge server. This is useful in areas like vision systems, environmental monitoring and logistics.

Smart infrastructure — Certain areas of smart infrastructure require low-latency processing that can be provided by edge computing systems and edge data centers. The most prominent application area is autonomous vehicles, where cars will need a mixture of compute capabilities as well as a connection back to the surrounding infrastructure. Applications like traffic monitoring and prediction, utility monitoring, and emergency services are some areas where edge computing can provide improved service delivery by interacting with vehicles and other infrastructure systems.

AI at the edge — Among all the high-compute tasks that can be performed on an embedded device, AI at the edge requires the greatest level of compute. Today’s chipsets and on-device model acceleration can be used to perform the lowest-compute tasks, while higher-compute inference and training requiring data from multiple systems can be performed at the edge. The cloud can then serve as an access point for data, system configuration and service monitoring.

Security and defense — Systems deployed in these areas are becoming progressively more complex and they require a significant amount of data to operate successfully. In some cases, network connectivity to the cloud may be intermittent or non-existent. Edge computing fills in the gap by providing on-demand compute resources and services like sensor monitoring, navigation and secure communications. Newer edge server systems for security and defense are also implementing AI applications as part of monitoring and threat detection.

5G and mobile services — Colocation at a base station is one approach to service provision for 5G networks, as well as upcoming 6G networks. These edge compute nodes enable content delivery, low-latency SaaS platforms and localized data processing in end-user applications. Edge computing plays another important role in relieving network congestion for high bandwidth data streams in certain embedded applications as data would not need to be sent to the cloud over the network backhaul.

Engineers should focus on the ecosystem

Splitting processing tasks between the end user’s device, an edge server or edge data center, and the cloud requires engineers to work at multiple levels in terms of application development. Cloud-enabled services require only two levels, with the highest compute tasks being offloaded to the data center. Additional application development is needed to implement the most time-critical portions of service delivery at the edge instead of in the cloud.

The concept of edge computing implies that companies would need to develop their own data centers, but this is not really the case. Edge computing is instead developing into a model resembling the standard cloud computing model, and colocation is available from some edge data center companies. At a high-level, development teams can identify and execute their edge computing business model by:

  • Identifying the most time-critical processing requirements, and developing an edge application to handle those tasks.
  • Identifying the data required most often by the end user, and planning to store these data at the edge or on user equipment.
  • Implementing the remaining processing tasks that are not critical for time-sensitive service delivery in a cloud application.

With high compute offloaded to the edge, systems designers can focus more on simplifying the design of client devices. Hardware designers can focus on cost optimization rather than increasing the complexity of the end user device with a higher compute chipset.

The tradeoff is lower latency in service delivery, but at the cost of more development due to the edge application acting as a middleman. Now, the ecosystem and the market for edge-enabled services and access to processing power is still growing. As the market matures, companies will be able to access these advanced services

To contact the author of this article, email engineering360editors@globalspec.com


Powered by CR4, the Engineering Community

Discussion – 0 comments

By posting a comment you confirm that you have read and accept our Posting Rules and Terms of Use.
Engineering Newsletter Signup
Get the Engineering360
Stay up to date on:
Features the top stories, latest news, charts, insights and more on the end-to-end electronics value chain.
Advertisement
Weekly Newsletter
Get news, research, and analysis
on the Electronics industry in your
inbox every week - for FREE
Sign up for our FREE eNewsletter
Advertisement
Find Free Electronics Datasheets
Advertisement