Cloud-Native Automotive Edge: AGL for Software-Defined Vehicles
summary
Proposes a concrete, working stack: VirtIO device virtualization as a standard interface that already supports block, input, GPU and network across hypervisors and cloud, plus planned audio video camera and CAN in the next three years. Demonstrates a unified virtual display built on virtual GPU that spans HVAC, instrument cluster, HUD and tablet and lets one app render across multiple screens. Runs identical AGL binaries on AWS Graviton (KVM) and on automotive SoC (OpenSynergy hypervisor) and even on a MacBook, and shows the demo functioning.
AI-written summary · about
chapters (13)
- 0:00Intro
- 2:03About AGL Virtualization Expert Group
- 2:34Cockpit Digital Transformation
- 4:04Device Virtualization: Specific Necessity in Automotive
- 7:22Device Virtualization: Key to Software Defined Vehicles
- 8:37Overview of Device Virtualization in AGL-Concept
- 9:14Pains around Virtualized AGL in the Past
- 10:02Enter Standard Virtualization Framework - VirtIO
- 11:46VirtIO Work for Hypervisor Environment
- 14:48VirtIO Work for Non-Hypervisor Environment
- 20:47Unified HMI Schematic Architecture
- 28:33Trends and Challenges driving the need for an SDV strategy
- 30:26What is "Software Defined"? How to make it possible?
related talks
30mEnabling a Software-Defined Automotive Edge with VirtIO Based Device Virtua...
27mDeep-Dive in VirtIO: Virtualized AGL with VirtIO to Achieve Cloud-Native Environment...
20mIntegrate RISC V to build Open Common Automotive Platform
17mTowards Intelligent Automobile Cockpit via A New Container Architecture
49mOpen Source in Every Car with Automotive Grade Linux
17mKeynote: Beyond the Cloud: Edge Computing
20mSOAFEE-Based Software Defined Vehicle Development for Autonomous Driving
49mOpen Source in Every Car with Automotive Grade Linux