Policy Driven Latency Control for Vehicular Real Time Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional vehicle communication networks face limitations in bandwidth, real-time challenges, and incompatibility among channels, making it difficult to integrate multicore gigahertz system on a chip processors with existing real-time legacy devices and non-real-time applications without substantial recoding, and ensuring security from hostile attacks.
Innovation Solution
A policy-driven latency control system using a real-time partitioning separation kernel on a multi-core processor, with hardware virtualization and secure shared memory, allocates clock, memory, and I/O resources to meet focused latency ranges for Observation, Decision, and Execution processes, ensuring secure and efficient communication between electronic control units.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional vehicle communication networks use standard protocols (CAN, LIN, Flexray, MOST), then compatibility and ease of operation are improved, but bandwidth and real-time performance deteriorate
Solution Approach 1:
The system divides the vehicle network into multiple isolated virtual networks (real-time critical, real-time non-critical, non-real-time) that can operate independently on the same physical Ethernet infrastructure, allowing each segment to optimize for its specific requirements while sharing the overall bandwidth
Solution Approach 2:
A single Ethernet physical network replaces multiple legacy protocols (CAN, LIN, Flexray, MOST), providing universal communication capability across all vehicle systems while maintaining protocol-specific virtual channels through the separation kernel
2Quantity of substance
If Ethernet is used for higher bandwidth, then quantity of substance (bandwidth) is improved, but real-time performance and reliability deteriorate due to jitter and lack of bandwidth reservation
Solution Approach 1:
The system segments Ethernet traffic into different virtual networks with guaranteed bandwidth allocation, reserving specific portions for real-time critical and real-time non-critical traffic to eliminate jitter and ensure predictable real-time performance
Solution Approach 2:
The separation kernel acts as an intermediary layer between the physical Ethernet network and application layers, providing bandwidth reservation, traffic prioritization, and real-time guarantees that pure Ethernet lacks, while still utilizing Ethernet's high bandwidth capabilities
3Reliability
If hardware virtualization is used to isolate guest domains, then reliability is improved through sandboxing, but device complexity increases due to separation kernel requirements
Solution Approach 1:
The system uses hardware virtualization to segment the operating system into isolated guest domains (real-time critical, real-time non-critical, non-real-time), providing reliable sandboxing while managing complexity through automated resource allocation policies
Solution Approach 2:
The separation kernel implements self-service mechanisms where virtual network interfaces automatically negotiate bandwidth allocation and latency requirements with the policy manager, reducing the complexity of manual configuration and management
4Productivity
If multicore gigahertz processors are integrated, then productivity is improved through increased computing power, but ease of operation deteriorates due to incompatibility with legacy standards and substantial recoding requirements
Solution Approach 1:
The separation kernel provides a universal compatibility layer that allows legacy real-time applications to run on modern multicore gigahertz processors without modification, while simultaneously enabling non-real-time applications and future-proof architectures
Solution Approach 2:
The separation kernel acts as an intermediary between legacy real-time applications and modern multicore processors, providing the necessary abstraction and adaptation layers to maintain compatibility without requiring substantial recoding of existing systems
Data Source
AI summary
A system includes a real-time partitioning separation kernel installed on a multi-core processor. Guest operating systems are hosted with in hardware virtualized machines in the cores. Another hardware virtualized machine performs a real-time USB-CAN interface communicatively coupled to distributed electronic control units which acquire data and command actuators. A plurality of hardware virtualized machines support processes of various criticality. A secure shared memory serves as the communication means between processes performing different levels of functionality at suitable latency ranges. a policy to distinguish, allocate, and distribute clock, memory, and input/output resources to meet focused latency ranges to the Observation, Decision, and Execution processes. Remaining resources have diffuse latency ranges made available to the Observation, Decision, and Execution processes in an as available but guarded minimum and maximum buffet. A latency policy ensures that each process receives its minimum tranche before queueing for up to the maximum at the resource buffet.