Re: [saag] draft-iab-crypto-alg-agility-00

[Stephen Kent <[e-mail]>]

Ben,

Sorry to jump on late on this, but as a co-author of an RFC that deals
with alg agility for the RPKI, I thought I might be able to contribute.

It's generally useful to have a MUST implement alg (or set, if there are 
multiple
algs needed) for any set of protocols (data formats, etc.) to help ensure
interoperability. Selecting the MUST implement is sometimes hard, but 
it's worth
the effort.

Then one needs to have a way for all clients, CA, and log servers (in 
the CT context)
to express which alg9s) are being used at any time, so that everyone 
else knows
what choices have been made, when choices are allowed, and to guide 
folks through the
transition process.

I suggest folks read RFC 6919 as an example of how that can be done, 
although
I'm not suggesting that CT needs to adopt as thorough a model as did the 
RPKI.

Steve