[Secdispatch] EDHOC

[Göran Selander <[e-mail]>]

Dear Secdispatch,

We have been advised to ask secdispatch to consider EDHOC: https://tools.ietf.org/html/draft-selander-ace-cose-ecdhe

Those that follow the ACE WG should be familiar with this draft. The problem statement and motivation for EDHOC is described in section 1. In brief, the target is a lightweight key exchange protocol suitable for IoT applications, which:
a) has small message size and reuses existing IoT primitives to enable low overhead and small code footprint; 
b) is not bound to a particular transport, to enable end-to-end security in IoT deployments with varying underlying layers; and
c) can be used to key OSCORE (draft-ietf-core-object-security) that is lacking a harmonizing key exchange protocol.

These requirements are motivated by constrained IoT device deployments, but the protocol is applicable to other end-to-end security settings where the overhead due to security needs to be low. EDHOC addresses these requirements and builds on the SIGMA construction for Diffie-Hellman key exchanges. EDHOC, like OSCORE, is built on CBOR (RFC 7049) and COSE (RFC 8152) and the protocol messages may be transported with CoAP (RFC 7252).  

There has been a number of reviews of different versions of the draft; both by people who want to deploy it and by people analysing the security. A formal verification was presented at SSR 2018. There are a few implementations of different versions of the draft. The ACE WG has expressed interest in this work in several f2f meetings.

Please let us know if some information is missing for secdispatch to consider this draft, or how we can help out in the process.

Best regards
Göran, John, Francesca