HIPRG meeting
Thursday December 6, 2007
1300-1500 Oak Room

Notes courtesy of Andrew McGregor.

Approximately 40 people attended the RG meeting.

Administrative note:
Supplemental site (wiki) was taken down due to spam, is moving into the IRTF  
wiki, details to be posted.

-Security issues with HIP middlebox traversal.
draft-heer-hip-middle-auth-00.txt

P2P WiFi sharing architecture talk/demo. (Tobias Heer)
--20 min

Philip Matthews: There has been a lot of work in the IETF on how to  
talk to middleboxes... 15 ways, only 3 non-failures.  You might want  
to go off and join the discussion so this doesn't disappear.

Miika Komu: I think this is a good application of HIP.

Andrew McGregor: I'm a bit concerned about making the responder solve  
a puzzle... that somewhat breaks the design of HIP.  Also, this could  
be a solution for authenticating to IPv6 firewalls to allow HIT-based  
ACLs.

Martin S: I'm somewhat confused by what the middlebox is... is it a  
NAT or what?

A: It could be a more generic middlebox.

Miika K: It could also be a MM proxy/rvs.

-HIP and the Secure Mobile Architecture (SMA)- (Richard Paine)
--10 minutes

Andrew M: I totally agree about the size of the SCADA market... it's  
huge.  Also, this is a really cool bit of networking :-)

Philip M: So, this is basically a HIP-based VPN?

A: Sort of... there's broader access control, in that sometimes these  
need to talk to, for example, logging services.

-Update on HIPL experimentation and InfraHIP II project. (Miika Komu).
--15 min

-Host Identity Specific Multicast (Andrei Gurtov for Rolland Vida).
--20 min

- P2PSIP-related presentations

David Bryan: It would be nice for some HIP people to show up in  
P2PSIP lists, and post impressions of these talks there too.

Miika: Would it be useful to have a HIP intro at the next IETF?

Dave: I think so, but it's a bit late for P2PSIP

Andrew M: I think the multicast material looks like a useful  
primitive.  Also, HIP has a fairly strong running-code tradition,  
which should be emphasised.

-P2PSIP proxy talk/demo. (Joakim Koskela)
draft-hautakorpi-p2psip-with-hip-01.txt
--15 min

?: I think there's a HIP draft which explains HIP resolution via DHT?

A: This is about the ability to send the HIP base exchange via the  
overlay, to help with NAT traversal for example.

Andrew M: General P2PSIP question: how much interest/work is there in  
P2PSIP about supporting applications other than SIP?

Philip M: I'd say a year ago, primarily SIP, but now there's much more  
interest in running general applications, and all the leading  
proposals do at least theoretically support more general apps.


-P2PSIP system that is using the NodeID architecture (Martin
Stiemerling). Experience on HIP deployment in Ambient Networks.
--20 min

-- this is a general overlay (I was a bit previous)

Tom H: How large a client population have you tried?

MS: Some, but not huge as so far we've largely been functional testing.

MS: zeroconf and bootstrap are part of the project... just not part of  
this talk

-Problems in using HIP for P2PSIP
(Philip Matthews)

Credentials: Why not just use a credential-bound HI?

Bare HIT: Overlay should do neighbour discovery for these

PM: node running in multiple overlays simultaneously
regarding HIT vs IPv6 LSI

PM: HIP allows P2P overlays to run legacy apps without change
discussion of applications using embedded addresses

AM: practical experience with HIP shows that it's not a very big problem

AM: another possibility is to have the overlays try and route the HIT