> Markdown version of [/videos/736-a-solution-to-embed-container-technologies-into-automotive-environments](https://www.wearedevelopers.com/videos/736-a-solution-to-embed-container-technologies-into-automotive-environments). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# A solution to embed container technologies into automotive environments

How do you run containers when safety regulations demand a two-second boot time? Discover how a custom Rust platform and cached configurations cut automotive startup times by 67%.

- **Speakers:** Falk Langer, [Lukas Stahlbock](https://www.wearedevelopers.com/@lukas-stahlbock)
- **Event:** World Congress 2023
- **Published:** October 6, 2023
- **Duration:** 30:33
- **URL:** https://www.wearedevelopers.com/videos/736-a-solution-to-embed-container-technologies-into-automotive-environments

## Summary

Traditional software DevOps practices often fail when applied directly to autonomous vehicles and embedded devices due to resource constraints, intermittent connectivity, and strict safety regulations. To bridge this gap, engineers must adopt a specialized methodology encompassing ML Ops, IoT Ops, and a "mission control" approach to manage software rollouts to these nomadic devices. While containers offer an established deployment standard, traditional orchestration frameworks introduce unacceptable overhead for embedded hardware. For example, strict startup time regulations—like the two-second requirement for rearview cameras—demand optimizations beyond standard container runtimes.

By recognizing that automotive updates occur via dedicated routines rather than dynamic reboots, teams can cache file system layers and OCI configurations. Embedding these cached configurations directly into the container runtime reduces container startup time by up to 67% with minimal execution latency, offering an ideal trade-off between isolation-based security and embedded performance constraints. However, lightweight orchestration tools like K3s and Rancher still proved too resource-intensive for automotive reference hardware, prompting the development of a custom, rust-based IoT management platform tailored specifically for vehicle resource limits.

This bespoke IoT management platform integrates seamlessly into an "AI in the loop" pipeline, enabling continuous data collection, model training, and containerized deployment. Real-world testing on robotic proxies demonstrates how sensor data recorded in MCAP formats can be extracted, annotated, and fed into localized, small-scale AI models. Ultimately, managing DevOps for vehicles at scale requires precise traceability between recorded environmental data and the resulting neural network weights, allowing manufacturers to safely orchestrate staged deployments across fleets of millions of vehicles.

**Keywords:** IoT device DevOps, ML Ops methodology, automotive containerization, container startup optimization, nomadic devices, OCI configuration caching, container runtime tuning, embedded system constraints, K3s orchestration overhead, rust-based device manager, AI in the loop, fleet management rollout, MCAP data annotation, real-time container latency, vehicle software traceability

## Chapters

1. **Challenges of applying DevOps to autonomous vehicles** (00:02) — Why standard software development cycles fail for self-driving cars and embedded systems.
1. **Expanding continuous integration to MLOps and IoTOps** (02:50) — Adapting iterative delivery loops for machine learning and disconnected nomadic devices via mission control.
1. **Resource constraints and non-functional requirements for containers** (06:58) — Overcoming resource limitations and execution latency rules when operating containers inside closed vehicle environments.
1. **Optimizing container startup times on embedded reference hardware** (08:53) — Caching container configurations and merging root file systems to bypass native toolchain overhead.
1. **Developing a lightweight custom IoT management platform** (14:31) — Replacing resource-heavy orchestration frameworks with a custom Rust-based device manager for embedded architectures.
1. **Executing AI in the loop on robotic test hardware** (18:08) — A demonstration of gathering data, training models, and deploying targeted updates back to autonomous robots.
1. **WebAssembly and staging deployments for large vehicle fleets** (25:09) — Addressing audience questions about runtime alternatives and phased update strategies for commercial automotive deployments.

## Related Moments

- [Tackling functional complexity with localized vehicle electronic architectures](https://www.wearedevelopers.com/videos/258-on-developing-smartphones-on-wheels) (from "On developing smartphones on wheels")
- [Evaluating automotive software architectures and backend technologies](https://www.wearedevelopers.com/videos/374-the-future-of-automotive-mobility-upcoming-e-e-architectures-v2x-and-its-challenges) (from "The future of automotive mobility: Upcoming E/E architectures, V2X and its challenges")
- [Practical learnings from deploying containerized AI solutions](https://www.wearedevelopers.com/videos/1127-using-containers-to-deploy-ai-models-across-our-microscopy-platform) (from "Using Containers to deploy AI Models across our microscopy platform")
- [Addressing edge management challenges in the automotive industry](https://www.wearedevelopers.com/videos/67-the-best-of-two-worlds-bringing-enterprise-grade-linux-to-the-vehicle) (from "The best of two worlds - Bringing enterprise-grade Linux to the vehicle")
- [Unified tech stack and hardware for Volkswagen](https://www.wearedevelopers.com/videos/519-finding-the-unknown-unknowns-intelligent-data-collection-for-autonomous-driving-development) (from "Finding the unknown unknowns: intelligent data collection for autonomous driving development")
- [Optimizing and deploying containerized AI inference workloads](https://www.wearedevelopers.com/videos/920-wwc24-ankit-patel-unlocking-the-future-breakthrough-application-performance-and-capabilities-with-nvidia) (from "WWC24 - Ankit Patel - Unlocking the Future Breakthrough Application Performance and Capabilities with NVIDIA")

## Related Articles

- [Building AI Solutions with Rust and Docker](https://www.wearedevelopers.com/magazine/494-building-ai-solutions-with-rust-and-docker)
- [How software is steering vehicle technology](https://www.wearedevelopers.com/magazine/515-how-software-is-steering-vehicle-technology)
- [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it)
- [MLops – Deploying, Maintaining And Evolving Machine Learning Models in Production](https://www.wearedevelopers.com/magazine/115-mlops-deploying-maintaining-and-evolving-machine-learning-models-in-production)

## Related Jobs

- [Software Engineer, Infrastructure Platform](https://www.wearedevelopers.com/jobs/ext/1940513-software-engineer-infrastructure-platform) at **Docker, Inc.**
- [Principal Software Engineer, AI Tools and Security](https://www.wearedevelopers.com/jobs/48450-principal-software-engineer-ai-tools-and-security) at **Docker, Inc.**
- [Software Engineer, Infrastructure Platform](https://www.wearedevelopers.com/jobs/ext/2004010-software-engineer-infrastructure-platform) at **Docker, Inc.**
- [ML Engineer](https://www.wearedevelopers.com/jobs/48448-ml-engineer) at **Docker, Inc.**
- [Staff ML Engineer](https://www.wearedevelopers.com/jobs/48463-staff-ml-engineer) at **Docker, Inc.**
- [Principal Solutions Architect, Professional Services](https://www.wearedevelopers.com/jobs/ext/2820116-principal-solutions-architect-professional-services) at **Docker, Inc.**