------------------------------------------
          IETF77
          CSI WG meeting agenda
          ------------------------------------------
          MONDAY, March 20, 2010
          1740-1940 Afternoon Session III
          ------------------------------------------
          
          Agenda bashing, note takers
          All - 5 min

Gunther on bakeoff: will try again next IETF in Maastricth          
 
          WG status
          Chairs - 5 min
          
Pretty much done with all deliverables except crypto agility.
 
Hash Analysis and  Proxy Send Problem statement - in IESG past WG
 
Cert work - thru LC
 
DHCP CGA interaction - WG LC to be issued soon
 
Only remaining work: PK agility
 
          Crypto Agility
          Tony - 15 min

Signature agility 
draft-cheneau-send-sig-agility and draft-cheneau-ecc-agility
 
ECC/ECDSA and advantages thereof, the usual ECC goodness
 
removed csi-cga-pk-agility - M-CGA was too complex
 
removed mention in sig-agility
 
hosts:
	* generate either RSA/SHA1 (legacy) or ECC/SHA256
	* ability to talk to RFC 3971 hosts
	* ability to verify certs other than that used   for the signature
 
router:
	* verify more than one signature type
	* different CGAs lead to different cert paths, behave as different entities
 
supported sig alg options (SSA) in ND message
	* per alg: sign and verify or verify only
	* ECC values follow IKEv2 registry values
 
universal signature option (USO)
	* compatible with RSA sign option when key id = 0 (legacy)
 
3971 nodes will ignore SSA and USO and just reply with their  legacy sig
 
new nodes can verify and sign diff signature families. 
	* or not, but it is clear and will respond accordingly
 
implementation  being done but need more expert review
 
          Implementation report
          Tony - 10 min
 
implementation within the MobiSEND project in France
 
modified SCAPY6 (tool for packet generation) for SeND
	* mostly in user space
	* must run as root
 
to be added:
	* sign agility support
	* CRL check
	* rate limiting
	*  in-kernel CGA generation
 
not production but proof-of-concept
 
Thomas: what is the diff between your and docomo's implementation?
 
Tony: This is a re-implementation. for improved deployability and stability as well as extensibility
 
          Next steps with crypto agility
          All - 15 min
 
energy problem in the group.
 
specially worrisome with this agility work
	* it would modify base spec
	* insufficient review
 
Thomas: would be good to try to obtain good review about this, or publish as experimental
 
Jari: if it has been tried and nothing has happened, it is better to let it expire, not to publish it
 
marcelo: if no change, we will close as soon as we finish up the other drafts. If there's energy we will continue.