Autonomous Vehicle Edge Computing for Distributed Data Center Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Centralized data centers face high bandwidth and cost burdens due to the need to service widely distributed users, which can be alleviated by leveraging the computing resources of autonomous driving vehicles (ADVs) when they are idle.
Innovation Solution
ADVs are repurposed to function as micro data centers by reconfiguring their software stack using over-the-air programming, allowing them to provide computing services to local users when not in use for driving, thereby distributing computing power closer to users and reducing infrastructure costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If data centers are centralized to save facility maintenance costs, then facility maintenance cost is reduced, but network bandwidth requirement and cost increase due to serving widely distributed users
Solution Approach 1:
The patent segments the centralized data center into distributed edge computing nodes deployed in autonomous vehicles. Instead of one large centralized facility, computing resources are divided and distributed across multiple vehicle-based nodes, allowing local service delivery and reducing network bandwidth requirements while maintaining cost efficiency
Solution Approach 2:
The patent introduces mobility as a new dimension to traditional data centers. By deploying computing resources in autonomous vehicles that can physically move to different locations, the system transforms from a static centralized model to a dynamic distributed model, simultaneously reducing facility maintenance costs and network bandwidth requirements
2Ease of manufacture
If data centers are centralized to reduce infrastructure costs, then infrastructure cost is reduced, but network cost increases due to high bandwidth requirements
Solution Approach 1:
The patent enables autonomous vehicles to provide computing services to local users independently, without requiring constant connection to centralized data centers. Each vehicle-based computing node serves its local community autonomously, reducing both infrastructure costs and network costs by eliminating the need for high-bandwidth connections to centralized facilities
3Ease of manufacture
If computing resources are concentrated in centralized data centers, then facility maintenance cost is reduced, but computing resource utilization efficiency decreases when vehicles are idle
Solution Approach 1:
The patent makes autonomous vehicles universal computing platforms that can perform both their primary autonomous driving function and secondary distributed computing function. When vehicles are idle after completing driving tasks, their computing resources are automatically allocated to provide cloud computing services, thereby increasing computing resource utilization efficiency without requiring additional centralized facilities
Data Source
AI summary
Embodiments are disclosed of a data center including a plurality of computers. A communication interface is communicatively coupled to the plurality of computers and adapted to communicate, wirelessly or by wire, between the plurality of computers, a user device, and at least one autonomous driving vehicle (ADV) computer. Instructions are stored in the plurality of computers that cause one or more of the plurality of computers to receive a user computing request from the user device. The computers determine whether there is an available ADV computer to execute the user computing request. If there is an available ADV computer to execute the user computing request, the computers reconfigure the available ADV computer from autonomous driving mode to computing mode, and transmit the user computing request for execution by the ADV computer. If no ADV computer is available to execute the computing request, then the request is executed by one or more of the plurality of computers.