Re: [Sidrops] 8210bis

[Martin Hoffmann <[e-mail]>]

Randy Bush wrote:
> 
>    There are zero or more 32-bit Provider Autonomous System Number
>    fields; see [I-D.ietf-sidrops-aspa-profile].

As a mere implementer, I can’t really judge how many Provider ASNs a
single ASPA typically will have, so I am just throwing this out there
for consideration:

ROAs are decomposed into a separate PDU for each prefix. The advantage
is that the PDUs are fixed size and I can keep read buffers on the
stack, whereas for variable length PDUs I will have to have some kind of
allocation strategy. Obviously, there is a 96 byte overhead for each
additional Provider ASN if we did the same for ASPAs. 

Am I assuming correctly that decomposing the ASPA would also make the
other paragraph you were asking for feedback for unnecessary:

>    The ASPA PDU is to support [I-D.ietf-sidrops-aspa-profile].  Receipt
>    of an ASPA PDU announcement when the router already has an ASPA PDU
>    with the same Customer Autonomous System Number, replaces the
>    previous one.  This is to avoid surprises when a BGP announcement is
>    received between an withdrawn PDU and a new announced PDU.

Kind regards,
Martin