Version 18 of ARC Specification Published

On October 2nd, the IETF’s DMARC Working Group published version 18 of the ARC specification document, incorporating the changes suggested during the Working Group Last Call (WGLC).  There are many small language changes to clarify concepts or bring usage in line with other IETF documents, such as Email Address Internationalization (EAI).

You can review the changes between the pre- and post-WGLC versions using this link.

Working Group Completes Last Call For ARC Specification

The DMARC Working Group has completed Last Call for the ARC specification. This means that when the final consensus changes are incorporated, the document will be submitted for approval and publication by the IETF. It is possible that the specification might be published before the next IETF meeting in November, though perhaps more likely between then and the following meeting in March 2019.

The ARC specification has been under development by the Internet Engineering Task Force’s (IETF) DMARC Working Group since roughly June 2016. In July, just over two years later, the Working Group reached a milestone known as Working Group Last Call (WGLC). During Last Call the members of the group determine if consensus has been reached on the content of the document in question. If so, any final changes are made to the document and it is submitted to one or more IETF officials known as Area Directors. Upon review they may ask questions or make suggestions, and when satisfied, the document is approved for publication.

The ARC specification grew out of work begun by a group known as OAR in the summer of 2014. OAR refers to a proposed email header named Original Authentication Results, the subject of a draft proposal published in 2012.

New ARC Implementations at IETF 99 Hackathon

An ARC table was organized at the weekend hackathon held before the IETF 99 meeting in Prague this week. ARC is an official work item of the IETF DMARC Working Group, and will be discussed during the working group meeting Thursday morning.

Participants from two different vendors worked on new implementations of ARC, compatible with their software stacks. By the end of the weekend, one of these implementations was able to validated multiple-hop ARC chains, which is perhaps the hardest part of any implementation. And a participant from a third vendor indicated they have made significant progress in developing their own ARC implementation prior to the weekend.

If you compare this to all proprietary and open source implementations known before the hackathon, it will represent a 60% increase in the number of ARC implementations when they are completed. In addition to the implementations mentioned, the weekend included more interoperability testing, bug hunting, patch generation, and pull requests.

Many thanks to all the participants at the hackathon, whether they were in the room or working remotely!