Civil Infrastructure Platform: Industrial Grade Open Source Base-Layer Development

### Overview In this presentation, Yoshitake Kobayashi (Toshiba) and Urs Gleim (Siemens) present an update on the Civil Infrastructure Platform (CIP), a collaborative Linux Foundation project founded in April 2016. CIP focuses on developing an industrial-grade, open-source base layer for embedded devices operating within civil infrastructure systems, including transportation, energy distribution, industrial automation, healthcare, and building automation. Unlike enterprise or consumer software, civil infrastructure equipment requires extreme operational lifespans—ranging from 10 to over 40 years—where rapid kernel churn is impractical due to stringent reliability, security, and safety certification requirements.

### Key Topics - **Long-Term Support Requirements:** Industrial equipment lifecycles involve 3–5 years of development, 2–4 years of regional customization, and decades in the field. Maintaining fragmented, proprietary Linux builds across hundreds of industrial IoT gateways and edge devices creates unsustainable security and engineering maintenance overhead. - **Collaborative Base Layer:** Rather than building a full GNU/Linux distribution, CIP focuses bottom-up on providing a shared, conservative base layer. Member companies include Siemens, Toshiba, Hitachi, Renesas, Codethink, and Plat'Home. - **CIP Core and Debian Collaboration:** CIP selected Debian as its primary reference distribution to leverage its package ecosystem and open-source license compliance framework (DEP-5 / SPDX). While Debian LTS provides five years of support, CIP aims to bridge the gap to 10+ years. Tooling explorations to generate minimal CIP Core system images include the meta-debian layer (deby), ISAR, and ELBE using BitBake and Debian binary/source packaging.

### Technical Details - **Super Long-Term Support Kernel (SLTS):** CIP selected Linux 4.4 for its initial SLTS kernel, supported for more than 10 years, led by Debian kernel maintainer Ben Hutchings. Kernel maintenance follows an "upstream-first" rule to avoid out-of-tree regressions, backporting only critical bug fixes and essential security features. CIP maintainers also actively assist Greg Kroah-Hartman by reviewing patches submitted to the stable 4.4 branch (e.g., catching memory leaks). - **Real-Time Linux (PREEMPT_RT):** Real-time capability is critical for industrial controllers. CIP joined the OSADL Real-Time Linux project as a gold member, integrating stable RT patches into CIP kernel branches. - **CIP Testing & Board at Desk:** CIP released the "Board at Desk 1.0" testing environment, utilizing open-source tools like LAVA and KernelCI. It enables individual developers and distributed teams to run automated testing on physical boards (e.g., Renesas boards, BeagleBone Black) and share regression results across projects.

### Takeaways CIP addresses long-term software sustainability for critical infrastructure by establishing a standardized, upstream-aligned foundation comprising an SLTS kernel, PREEMPT_RT, automated testing, and a minimal set of Debian-derived core packages. During Q&A, the speakers noted that toolchain reproducibility is currently managed via archived VM and container images, while upper-layer connectivity stacks, fleet update systems, and Y2038 issues remain recognized industry challenges planned for gradual integration into the project's roadmap.