The long-term path to a better SCION specification

SCION is an Internet path-aware technology coming from extensive research at ETH Zurich and other universities. The technology slipped out of the lab and started gaining its first productive deployments, powering networks such as the Swiss finance’s SSFN. Multiple protocol implementations then started to appear, both open source and commercial. With growing adoption, within the SCION community, there were stronger calls for standardization.

THE ROLE oF STANDARDS AND SPECIFICATIONS

The Internet is built on top of technical standards that are developed collaboratively. They ensure consistency, compatibility, and interoperability. The main reasons to standardize a technology are:

1) Better technology through peer review

Standardization brings the benefit of review from experts in the field. This collaborative process helps refine and improve the protocol, ensuring that the final specification is robust, secure, and efficient.

2) Interoperability: Speaking the same technical language

A well established and precise specification ensures that all implementations “speak the same language” and can effectively understand and communicate with each other. In addition, the protocol should be able to reuse and interoperate with existing networks as much as possible. This is crucial as we hope to see more and more vendors implementing SCION.

3) Openness: Building an inclusive ecosystem

For those looking to join the SCION ecosystem and develop products based on this technology, having access to comprehensive technical documentation is vital. A well-defined specification provides the necessary information, making it easier for new implementers to adopt SCION.

The role of the Internet Engineering Task Force (IETF)

To achieve an appropriate level of review for SCION, it is essential to engage with a global Standards Developing Organization (SDO). The Internet Engineering Task Force (IETF) is the prime SDO for Internet protocols, and we share the vision of “making the internet work better”. We therefore decided, together with our community of early adopters, to engage more and more at the IETF and its research branch, the IRTF.

Such engagement requires continuous effort and resources to develop specifications, attend discussions, and attend quarterly meetings. This often aligns badly with the situation of academics, and of some of our early adopters in various industries. We, as SCION Association, therefore invest significant resources to bridge the gap and bring the voice of our members to the IETF/IRTF. Our Standardization Committee gives input on the overall direction, and since its creation in 2023, decided to engage as much as possible with the IETF and IRTF.

The IETF Process

The IETF process is a collaborative, consensus-driven approach that involves the development, review, and approval of Internet standards. It begins with the creation of a working document called an Internet Draft (I-D) which is discussed and refined within working groups or research groups composed of industry experts and stakeholders focusing on particular areas of Internet technology, so it’s important to identity the appropriate group and submission track within the IETF. These working groups operate transparently with open meetings and public mailing lists to ensure broad participation, and ensure that a draft meets the necessary technical standards and interoperability requirements.

Once a draft achieves consensus, it moves through several stages of review depending on the track. After addressing feedback and achieving final approval, the draft is published as a Request for Comments (RFC). This rigorous and inclusive process ensures that protocols are robust, interoperable, and meet the diverse needs of the global Internet community.

DOCUMENTING THE CURRENT SPECIFICATION – ISE SUBMISSIONS

Following various rounds of discussions with IETF working groups and IRTF research groups, we decided to first document the core SCION protocol specification as it is currently deployed today, before working on protocol evolution. Since SCION was initially developed outside of the IETF by ETH Zurich, with significant contributions from our member Anapaya Systems, and as it falls within the scope of several areas, we decided to submit three Internet Drafts specifying the PKI, control plane and data plane components to the IETF ISE (Independent Submission Editor) stream. This stream does not involve going through the IETF consensus process, and does not produce standards. It is rather meant to introduce new ideas to the IETF community, while bridging between academia and engineering. In parallel, we’ll continue to engage in discussions with the community and intend to submit further drafts at the Path-Aware Networking Research Group (PANRG) discussing deployment experiences from our early adopters and open research questions around SCION.

Learning from early adopters

Research around SCION started with a clean-slate approach to make a better, more secure, and more performant Internet. Thanks to this vision, established by its inventor Prof. Adrian Perrig, the technology achieved unique security and availability properties, and gained its early adopters. When bringing such new technology to the field, there are often unexpected operational and practical issues that require bridges between the security research and practice. Based on these experiences, the technology needs to mature and evolve, so that it can be applied in more and more use cases, and become easier and simpler to use.

NEXT STEPS

The journey of SCION standardization is only at its beginning. The current Independent Stream documents are not IETF standards, rather they represent a solid starting point documenting how SCION works today. In the long term, we will need to use these specifications and the lessons learned to evolve the protocol to simplify it, further improve interoperability with existing protocols. This is why we continue discussions with the IRTF PANRG (Path-aware networking Research Group) and other working groups, with the hope of using this work as a basis for future work in the IETF.