OGF27 FVGA-WG Session Meeting Notes 
===================================
by Inder Monga and Ralph Niederberger

Ralph presented the agenda for this session, gave a short overview about the 
objectives of the working group, and presented the current draft protocol 
proposal (Firewall Virtualization-service draft version 2).

After describing the current status of work we started discussion on comments 
which came up during Ralph's presentation.

Comment by Ralph himself: No IETF comments received so far after doing much effort.  
Proposal: Sit together with AD and OGF IETF Liason - Cees De Laat.

Comment by Yuri Demchenko: You could use application secure session credentials for 
authorization and authentication rather than depending on the FiTP's secure 
session setup. These options should be supported to offer the most flexibility 
to the application.
Proposal: Look into the draft protocol description and provide input how to 
change the protocol draft for supporting these options. Provide this input via
the groups mailing list for further discussion.  

Comment by Richard Hughes-Jones (Infrastructure AD): Even though there are no 
comments to the protocol draft by IETF and others, we can consider making this a 
draft standard and then write an experience document after some time (1 year or so) 
with two independent interoperable implementations. We should look at taking 
this forward. 
Answer: The co-chairs felt this is too early to make it standard, but will discuss 
this offline.

Comment: if there is a well know service and the compute nodes are in the back, 
do you really need this service?
Answer: This depends on the application. If the application uses only well known ports,
which should be open all the time, it does not make any sense to use the FiTP protocol.
Then a predefined access rule would be recommended. But if those ports really need to be
opened only for time frames unknown in advance and ports/port ranges may differ from 
application to application, the FiTP protcol would make sense. So usage will be 
application driven/dependent. A usage scenario document will be provided in future.     

At the end of the session the session chairs asked the audience again to read and comment 
on the current document. Any input would be very welcomed.