OSPF WG Meeting - Chicago - July 24, 2007 
    Scribe: Michael Barnes (mjbarnes@cisco.com)

13:22 Review agenda - Acee Lindem

13:25 Review working group progress - Abhay Roy

    OSPFv3 MTR draft needs more discussion
    OSPFv2 HMAC-SHA1 draft needs more discussion

13:30 OSPFv3 MIB - Acee presenting slides for Dan Joyal

    Want to do WG Last Call soon

13:35 Draft Authentication/Confidentiality for OSPFv2- Suresh Melam

Acee: 
    A single hop virtual-link packets could have the same src/dst as non-VL 
    packets on the link.

Vishwas Manral:  
    SPD is defined by IPSec, but is internal to an
    implementation so not limited to what RFC specifies. But
    implementations don't support it.

Bill Atwood:
    Work being presented in routing security WG for
    automatic  rekeying. Work going on in PIM for rekeying to protect
    link-local addressing. The last slide references the automatic
    rekeying, it should be linked to the other work.

Acee: Straw poll as to whether to make WG document, some support, a 
         couple of folks against. Taking it to the list.

13:54 Extensions to OSPF for Inter-AS TE - Mach Chen

Dave Ward:
    Asking if the working groups feels a new link type is required

Acee:
    Doesn't seem like this is a link

Kireeti Kompella:
    Should just be a TE LSA. Do not have adj or hellos, but want to add into
    database to do calculations.

Dave Ward:
   Do we need to carry the AS number and what is it used for?

Adrian:
   Can be used for signalling (details lost here) 

Kireeti:
   Useful to carry the Router ID, ambivilent about the AS number

Dave:
   Would like to see why, where, when and applicability for carrying the 
   AS number.  Router ID is sufficient but necessary. AS number is not necessary

Adrian:
   From CCAMP: close to last calling it.

Acee:
    Seems like this is a connection to an AS as opposed to a link. Do we really
    want to call it a link, or should be a top level TLV? Two top levels now:  
        Link TLV and Router Address TLV.

Mach:
   ?

Acee:
    So it is tied to the link.

Abhay:
   The link in a TE LSA are additional attributes to the physical link.

Kireeti
   There is a link, it's passive

Acee:
    AS Number is passive information

Kireeti:
    Since no adjacency want full TE info, can't do calculation if don't have
    A2-B1 link

    Would like explicit statement saying no hellos and no adjacency. 
    Draft implies that. Do not want an inter-AS adjacency.

Abhay:
   Need a new link type?

Kireeti:
   Could be point to point

Dave:
   Doesn't matter because it's passive

Kireeti:
   Don't think a new type is needed.

Mach:
   Need a new type to indicate that is inter-AS

Adrian:
   Point was that inter-AS TE how careful do you want to be to avoid confusion,
   A new type avoids any ambiguity. If not worried about ambiguity a new
   type may not be needed


14:15 OSPF Charter Update - Acee

Decided to make an incremental update this time around.

Dave W: 
    Solution to separate routing from non-routing information

Abhay: 
    Okay to solve in v3 only?

Dave: 
   Advise to solve in v2 since it's so widely deployed
   Convergence is highest priority, so need a way to prioritize routing
   over non-routing.

Acee: 
    Reason?

Dave:
    Protect, preserve fastest possible convergence time.
    It's something that's been worked on in ISIS for a while now.

Acee:
    Multiple instaces supported by OSPFv3 

Dave:
    Think we need to have a solution for OSPFv2, not sufficient to require
    changing to OSPFv3.