[Idr] Re: I-D Action: draft-ietf-idr-deprecate-as-set-confed-set-15.txt

[Jeffrey Haas <[e-mail]>]

> On Aug 29, 2024, at 05:22, Claudio Jeker <[e-mail]> wrote:
>> 
>> The likely answer there is "bgp-5".
> 
> This sounds like the answer is "IPv6".

If you like. 

Some “issues” require a reset of baseline expectations. What we have today works. 

Such a reset is usually carried as part of other larger work. 

>> Go forth and do so in your stack. :-)
> 
> I think this comment is not helpful.

It wasn’t intended to be. 

Implementors are free to take unilateral actions they might think helpful to their users. Your request effectively was. “I don’t like this, what bludgeon can we use to force everyone to remove this?”


> 
> Just to be clear our stack enables 4-byte AS support by default on all
> session since 2009.

This is a reasonable default. 

 
> Lately we added a config option to enforce it and
> at some point I will make the enforcement the default and have users
> explicitly disable 4-byte AS support for those last few sessions that
> need it.

Enforcing it is a valid implementation choice. 


> In most cases sessions negotiate down to 2-byte for no good
> reason apart from bad config settings.
> 
> What are your stacks doing? Isn't it time to follow our footsteps?

For any feature where the “answer” is “change a long standing default or prevent something from working” it is often only a few lines of code change. 

At a larger vendor with an established user base, it’s also weeks of fights with the rest of the layers of the company who may be negatively impacted by “doing the right thing”. 

If your stack and your governance model have the freedom to make such changes freely… go forth and do them. 

Jeff


> 
> --
> :wq Claudio