How Distributed Edge Computing Leverages Best Practices for the Automotive Industry
By AECC President and Chairperson Ken-ichi Murata-san, AECC Directors Christer Boberg and Roger Berg
The AECC’s mission is to help automotive industry stakeholders prepare for the global deployment of fully connected vehicles. This 21st Century phenomenon is quite exciting, but there is a substantial amount of “under the hood” technology development required to provide a networking and computing infrastructure capable of supporting a global scale. Our bedrock idea is for the adoption of Distributed Edge Computing. While this idea is relatively new for the automotive industry, stakeholders can take comfort in that our model is grounded in decades of analogous use by many other industries.
The idea of Distributed Edge Computing is quite straightforward. It entails decentralizing the cloud and moving to compute processing closer to the source of data. Our general use case is for hundreds of millions of connected vehicles needing to gather, process and manage data for real-time connected vehicle services such as dynamic HD mapping.
Managing an ever-increasing volume of data is the deal-breaker. If the network and computing systems that process the data are not fully adapted to accommodate the uplink data uptake, it is impossible to ensure services are effective across millions of vehicles. Hence, Distributed Edge Computing, which places data processing closer to the vehicles, cuts the volume of data transmitted, enabling the ecosystem to deploy and add newer, next-generation service offerings.
Short History of Edge Computing
In the early years of cloud computing (1990s), users quickly discovered that retrieving and viewing large files of images and videos was slow as molasses. In response, the first iteration of edge computing was debuted by MIT researchers as Akamai with its content delivery network (CDN). The CDN used nodes of cached static content placed close to the end user for fast retrieval. Today, we couldn’t get our Netflix fix without this type of edge computing technology.
Today’s incremental advances in edge computing are now bringing fruits across many industries, such as connected vehicle services like dynamic HD mapping, which use data to present information in layers. The base layer is the mapping layer, which holds an HD map showing curb-to-curb information with centimeter accuracy from road markings and geographical boundaries. However, advanced dynamic HD mapping will show drivers images of changing road conditions from objects accidentally falling off a car just ahead to potholes or other road hazards.
Edge Computing for Connected Vehicles
In 2017, the AECC was created to coordinate and conduct development in use cases for connected vehicles. In particular, the AECC has also aimed to provide stakeholders with a reference architecture for network and compute infrastructure to support connected vehicles at a global scale. Stakeholder industries include automotive; mobile communications; big data; cloud and analytics; and application and service providers. Our intention is to also provide guidance toward the development of network and computing standards to support use cases for stakeholder industries’ business objectives.
The mobile, distributed nature of connected vehicles poses a perfect specialized use case for edge computing.
AECC’s Model for Distributed Edge Computing
To solve the problems of data processing and traffic on existing mobile communication and cloud systems, the AECC has introduced “Distributed Computing on Localized Networks.” The model entails several topology-aware localized networks to accommodate the connectivity of vehicles in their respective areas of coverage. Computation power is added to these localized networks to enable them to process local data, allowing connected vehicles to obtain responses in a timely fashion.
The concept is characterized by three key elements:
- Localized Network. A local network that covers a limited number of connected vehicles in a certain area. This splits the huge amount of data traffic into reasonable volumes per area of data traffic between vehicles and the cloud.
- Distributed Computing. Computation resources are geographically distributed within the vicinity of the localized networks’ terminations. This reduces the concentration of computation and shortens the processing time needed to conclude a transaction with a connected vehicle.
- Local Data Integration Platform. Integration of local data by utilizing the combination of the localized network and distributed computation. By narrowing relevant information down to a specific area, data can be rapidly processed to integrate information and notify connected vehicles in real-time.
Preparing for Connected Vehicles Everywhere
As you can see, the AECC is moving compute processing closer to the source of data. However, applying distributed computing to a mobile communication network is something new in recent years as it requires multi-industry integration and is not just copy and paste effort from a technical specification perspective. The architecture of AECC’s model and related impact on networking and computing infrastructures aims to solve high volume data requirements that are specific to automotive and mobile services use cases.
Our model also strives to work with existing global standards. As some standards may need iterations to meet new requirements for automotive use cases, the AECC is working closely with industry associations and standards bodies like 3GPP and IIC to help provide insights for successful outcomes. Our focus on use cases will also help achieve successful business outcomes.
To learn more, we invite you to read our General Principle and Vision white paper, and our recently-updated Technical Report, Driving Data to the Edge: The Challenge of Traffic Distribution. We also welcome collaboration by industry stakeholders as members of AECC and its working groups. More information is available on our website, https://aecc.org.