On 09/02/2014 01:23 AM, Hannes Gredler wrote:
> hi tony,

Fuehrt' Di' Hannes ;-)
>
> requiring send-side ordering is certainly intended well,
> but that in itself does not guarantee ordered delivery
> at the receiving end :-/.

I think I pointed that out already below. full ack.
>
> IOW in case you have an application that needs both
> the 1) prefix-LSA and the 2) extended-prefix-LSA you need
> to have the logic that checks presence of 1)
> at the receiving end anyway.
>
> rather than doing things on the send-side let me suggest
> to check if you have got your act together during
> route-computation time. (this is a bit akin to
> IS-IS fragment zero handling - ala - if frag zero
> is not present then all non-zero
> frag content has to be disregarded).

That is required anyway, the ordering was meant as suggestion for
implementors but given what Peter/Acee write (extented can be sent
without the normal LSA, this makes no sense apparently).

--- tony

_______________________________________________
OSPF mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ospf

Reply via email to