At Thu, 31 Mar 2011 17:02:53 +0000, Andy Newton wrote:
>
> On Mar 31, 2011, at 11:47 AM, Spies, Sebastian Martin wrote:
>
> > The AFI of ROAIPAddress (defined in 3. The ROA eContent of
> > draft-ietf-sidr-roa-format) is of type OCTET STRING (SIZE
> > (2..3)). For which reason is it necessary to have a variable AFI
> > size? We are having just two values 0001 and 0002 and it is
> > unlikely to see new address families at a count that exceeds
> > 2^16. So why should there be a third byte?
>
> I believe this is represented the same way as it is in RFC 3779. Section
> 2.3.3:
>
> IPAddressFamily ::= SEQUENCE { -- AFI & optional SAFI --
> addressFamily OCTET STRING (SIZE (2..3)),
> ipAddressChoice IPAddressChoice }
Yes, that byte is the optional SAFI. At least, that's how all the
running code of which I'm aware interprets it. At present we're only
certifying AFIs 1 and 2 with null SAFI, hence the text in section 3.3.
The text should say that the third byte is the SAFI. Good catch.
_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr
