[LACNIC/Seguridad] [lacnog] Bogon route objects in the LACNIC IRR

Ronald F. Guilmette rfg en tristatelogic.com
Jue Ago 19 11:33:52 -03 2021

In message <CAJBrruiDNGXLGLrS+=PuqKyq8z7vB1m5oa5edshEB=G1cRcg7Q en mail.gmail.com>, 
Roque Gagliano <rgaglian en gmail.com> wrote:

>If you really want to change how the ROAs work, please go ahead and submit
>an IETF draft and follow the process. All your points were discussed a lot
>during the design process.

I'm sorry.  It seems that I failed to make myself clear.

I have no interest whatsoever in making *any* changes at all to the existing
RPKI system *or* to any of its associated ROAs, either currently existing
ones or future ones.  I have not proposed that, I have not suggested that,
I have not even vaguely insinuated that, and indeed I would be opposed to that,
absent the kind of formal IETF process that you mentioned.

No  I have spoken only of the "old world" of RIR IRRs... the old system under
which the five Regional Internet Registries publish *non-cryptographic* route
objects themselves, via their respective Internet Route Registries.  (I thought
that I had made this clear, but apparently not.  In any case, I hope that I have
made it clear now.)

It's just the old RIR IRRs that concern me.  It has always been my default
assumption that the new world of RPKI ROAs can and will take care of itself.

With respect to the RIR IRRs... which are, in general, older than RPKI, except
in the case of LACNIC, it seems... those old IRRs contain or have contained a
lot of what I would call "long abandoned bogon garbage".  That's the only stuff
that concerns me.

Unfortunately, someone at some time within LACNIC made the decision to import
the entire universe of LACNIC RPKI ROAs, blindly, and en mass, into the old
style LACNIC IRR -without- first filtering out what I call the "bogon garbage".

As I have previously said, I think that this could be easily rectified by LACNIC
staff, and *without* just ignoring all of the thousands or tens of thousands
of perfectly legitimate RKPI ROAs.  Those could all and should all still be
imported on a route basis into the lACNIC IRR.  I have no problems with that,
and in fact I think it would be Good if this continued to happen.  The routes
that involve bogon AS numbers are a different matter however.  Those should not
have old-style route objects auto-generated for them which are then placed into

I'm sure that some people (Job?) may say "Well, just don't worry about it!  Bogon
route objects within the IRRs of the various RIR are not actually Bad and are
not really a problem." but I don't think such a claim will stand up to careful
scrutiny.  If bogon route objects within RIR IRRs are not "Bad" or at least
undesirable, when why is it the case that all four of the other Regional
Internet Registries (ARIN, RIPE, APNIC, AFRINIC) have already taken steps to
remove some or all of -their- bogon route objects from -their- respective IRRs?


Más información sobre la lista de distribución Seguridad