RE: comment on draft-arberg-pppoe-mtu-gt1492-02.txt
Paul Jakma <[email protected]>
| Newsgroups | gmane.ietf.pppext |
|---|---|
| Message-ID | <[email protected]> |
On Thu, 24 Nov 2005, Peter Arberg wrote:
> This was discussed initially, as it was also our first thinking,
> but for optimization purpose it was suggested to include the max.
> payload number as well.
I did try look over the previous discussions, I read the thread where
it was suggested to introduce new tags, I didn't see any discussion
of including a payload size though - never mind the rationale for
doing so.
> The idea of the new option was to have the PPPoE client communicate
> in the PADI and PADR the maximum MTU/MRU it can deal with. Then the
> PPPoE server will have a good hint to immediately select a proper
> value for the MRU to propose in its LCP config_req. This would make
> the regular case more streamlined.
Right, I did read the RFC. ;)
It doesn't give any rationale for why it's a good idea to include a
"Max-Payload" value in the tag though.
> It was not the intention to complicate things, simply to make the
> negotiation more streamlined. And when the TAG had to be included,
> it did not seem like an issue to include the max MxU number.
It's an additional piece of state for:
a) implementations to have to track and hence get wrong
b) future refinements to have to bear in mind wrt
backward-compatibility.
Making it be just a "remove the RFC2561 limit" flag simply reduces
things back to normal PPP, at least wrt MRU/MTU negotiation.
> The number is only to be set by the PPPoE client side, never from
> the server (BRAS) side, the PPPoE server will echo the clients
> value back as a okay lets do PPP MRU negotiation to settle on a
> value.
Why can't the client just start doing MRU/MTU negotiation as part of
PPP setup, as it will have to do anyway? Why add an additional value
into the mix?
This Max-Payload value seems quite redundant.
regards,
--
Paul Jakma [email protected] [email protected] Key ID: 64A2FF6A
Fortune:
"The medium is the message."
-- Marshall McLuhan
_______________________________________________
Pppext mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/pppext