Derek,
On 28/06/2021 19:46, Derek Fawcus wrote:
On Mon, Jun 28, 2021 at 09:39:34AM +0100, Bob Briscoe wrote:
My attempt to generalize this UK-specific knowledge: No-one is going to
add sequencing without a reason, and the only possible reason mentioned
so far would be stateful compression, which these days would just slow
things down more than the bandwidth saved.
I also mentioned PWE fragmentation - rfc4623 - it depends upon the sequence
numbers as a fragment identifier. So depending upon how it is implemented,
it may implicitly turn on sequencing, without the user knowing.
So the related question then becomes, where is that fragmentation support
being used - if anywhere?
[BB] Sorry, I omitted to explain that, in the PWE case, I didn't think
differential delay would be a problem, because all traffic in the PWE
would be expected to want the same low delay service. My concern was
about sequencing of tunnelled packets undoing deliberately
differentiated delay between different types of IP traffic within an
aggregate.
But thanks for picking me up on not explaining that.
Cheers
Bob
DF
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area
--
________________________________________________________________
Bob Briscoe http://bobbriscoe.net/
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area
