Skip to content

Latest commit

 

History

History
44 lines (35 loc) · 2.13 KB

SignatureFormatAlignment.md

File metadata and controls

44 lines (35 loc) · 2.13 KB

Aligning Signature Formats

by Manu Sporny, Christopher Allen, Jason Law, and Ryan Shea

As the Verifiable Claims work proceeds, it is important to ensure that all of the signature formats that are used by Verifiable Claims continue to follow a coherent design approach. Recent discussions at IETF 98 and WWW2017 have resulted in a breakthrough related to the design approach and standardization timeline related to Linked Data Signatures. The Rebooting Web of Trust community should discuss how their use of the signature formats in Verifiable Claims could benefit from the design changes outlined in the rest of this paper.

JOSE JWS and Linked Data Signatures

One of the reasons that Linked Data Signatures, which are used by Verifiable Claims, did not initially reuse the JSON Web Signing (JWS) specification was due to a variety of issues outlined in Analysis of JWS and Linked Data Signatures[1]. We met with Mike Jones and John Bradly, who were the core designers for JWS, at IETF 98 and worked together to align the needs of Verifiable Claims and Linked Data Signatures with what the JOSE stack is capable of providing. In doing so, we found a way that Linked Data Signatures can re-use JWS to achieve its goals by using a profile of JSON Web Signatures.

This new approach has the advantage of almost entirely aligning the Linked Data Signatures specification and the RSA Signature Suite with the JOSE stack, thus making standardizing the approach far less contentious than the previous approach.

Koblitz Curve and Linked Data Signatures

The Koblitz Signature Suite is not capable of re-using the JOSE work because JWS does not support Koblitz curve signatures. A deep analysis by the Crypto Forum Research Group does not exist for the secp256k1 curve. The signature suite can continue to use the current approach listed in the specification, but we may want to use this opportunity to discuss how the secp256k1 analysis could progress at the IETF's CRFG.

Next Steps

We desire to discuss this approach with other organizations that rely on digital signatures for their products and determine if this new approach is acceptable to those organizations.