>At the Sidr meeting at IETF 69, the following issue was raised in 
>regards to matching ROAs with NLRI in a BGP update.
>
>Consider an ISP that receives an allocation of a /20 prefix from source 
>A, and an allocation of an adajecnt /20 prefix from source B, but wants 
>to issue an advertisement for the aggregate /19 prefix.

I believe that there was another wrinkle to the issue.  (Since I was
the one who raised it, I know what I *meant*, although others might have
heard differently.)

The issue is that some RIRs, when allocating a prefix, will reserve a
matching adjacent prefix for future requests, so that the two
allocations are aggregatable.  This helps limit routing table bloat.
Note the difference: this is a *single* source allocating multiple
aggregatable allocations.  And I believe that it happens frequently,
by deliberate process.  

(Actually, I'm having trouble visualizing allocations from *multiple*
sources that are aggregatable - perhaps this only happens when ranges
rather than prefixes are involved?  I'm lazy and I don't want to do
the math - if you have an example of *prefixes* allocated from
multiple sources that were aggregatable, could you describe it?)

An aggregated cert from a single source does not run afoul of the
encompassing rules of 3779 and the res-certs draft, since it is from a
single source.  But the architecture notes that an RIR (or a source
further down the allocation chain) might not always want to create a
cert for the aggregate, as the validity times of the two longer prefix
allocations might be different.  Hence the need for multiple certs,
and the choice between either interpreting the validity of a BGP UPDATE
based on multiple (singly signed) ROAs or multiply signed ROAs.

I believe that multiply signed ROAs are a direct indication of the
intent of the prefix holder, while interpreting UPDATEs based on
multiple singly signed ROAs is trying to derive the intent of the
prefix holder.

I am in favor of multiple signatures on the ROAs.


--Sandy

_______________________________________________
Sidr mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/sidr

Reply via email to