[db-wg] RIPE NONAUTH route(6) objects using unregistered space cleanup - deployment *today*

Hi Ronald,

> On 29 Jun 2021, at 22:56, Ronald F. Guilmette via db-wg <db-wg at ripe.net> wrote:
> 
> In message <4A061BA5-46D6-4D42-BED1-663C3D498207 at ripe.net>, 
> Edward Shryane <eshryane at ripe.net> wrote:
> 
>> * The maintainer can ask to make an exception, not to delete a {bogon}
>> route(6) object.
> 
> OK.  I'll bite.  Why?

We don't know the purposes those route(6) objects are being used for. This is user data, and we can only delete objects according to the implementation plan agreed with the DB-WG.

We need an option to exclude deleting route(6) objects, in case it could cause an operational issue by deleting them automatically.

An alternative (if operational issues arise) is to disable the feature entirely, and this is also undesirable (i.e. we don't cleanup any objects).

> 
>> * Otherwise, after a further 2 weeks, delete the route(6) objects.
> 
> Wait a minute.  So for periods up to two full weeks, RIPE NCC will
> support and facilitate the use of bogon IP address space??
> 
> Somebody is going to need to explain to me the wisdom of that.

We will wait for two weeks after informing the object maintainer(s) to give them time to prepare, or in case they have a question, or raise a valid reason why we shouldn't delete the object.

The 2018-06 NONAUTH cleanup has a similar clause to properly inform the parties involved (14 days in the text):
https://www.ripe.net/participate/policies/proposals/2018-06

> 
>> Note for now, the AS origin status is *not* considered. Please discuss
>> further whether this should be included.
> 
> Maybe you missed it.  I have raised this point multiple times here on
> this mailing list, i.e. the issue of eliminating those route objects
> that make refernece to bogon AS numbers from the data base, and each
> time there was an outpouring of unbridled enthusiasm.  So what's the
> hold up?
> 

Your suggestion is being discussed and I await consensus from the DB-WG before proceeding with it (the feedback was not unanimous).

The current implementation for checking unregistered prefixes is already complete and can be extended to include AS numbers if the DB-WG agrees.

Regards
Ed Shryane
RIPE NCC


> Eliminating the route objects that referred to bogon IP address
> space was a GOOD and a NECESSARY thing, but failing to also eliminate
> from the data base all of the (mostly ancient and abandoned) route
> objects that refer to bogon ASNs makes about as much sense as using
> a colander as a condom.
> 
> 
> Regards,
> rfg
>