On Jun 11, 2014, at 12:00 PM, Saku Ytti wrote:

> On (2014-06-11 09:03 +0000), Vitkovský Adam wrote:
> 
>> However to be honest I'd like to see the existing protocols to scale better. 
>> In my opinion any new IGP protocol would get a cold welcome in carrier grade 
>> environments even though it could heal all the problems. 
> 
> It's quite unfortunate already today how ISIS and OSPF get features at
> different times as every feature needs to be specifically encoded to both
> protocols via separate RFC.
> I think this problem could be abstracted away by having common IGP container,
> then per IGP specification how container is encoded to wire in one specific
> IGP. Net result would be, that when new feature uses common IGP container
> format, it would be standardized for both/all IGPs at the same time.

given that OSPFv3 and IS-IS are now at par from an extensibility POV,
above would make a whole lot sense. i mean we do not need to
specify each and every bit redundant for both OSPF and IS-IS when all
what is need is an application object which is shared across
different protocol carriers. - the segment routing extensions are
actually an excellent example to demonstrate the IGP dilemma  …
the application level object is pretty much identical for IS-IS,
OSPF and BGP-LS, yet we need to I-D things three times …

/hannes
_______________________________________________
rtgwg mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/rtgwg

Reply via email to