Re: PPPoE: TLVs in Session?

Frank, 
 
The Credit TLV you mention isn't anything new --  it's part of RFC 4938.

 
Regards,
Stan


________________________________

	From: [email protected] [mailto:[email protected]]
On Behalf Of Frank Solensky
	Sent: Monday, September 08, 2008 1:20 PM
	To: [email protected]
	Subject: [Pppext] PPPoE: TLVs in Session?
	
	

	For those who haven't seen it yet, there's a proposed update to
RFC-4938: draft-bberry-rfc4938bis-00.txt.  One aspect of the revision
(section 3.3) includes adding a Credit TLV to the session protocol
packets.  This contradicts my understanding of the original PPPoE
specification in that the TLVs were present in the discovery packets but
the session packets always contain only the PPP packet as a payload.
However, in an off-list conversation with the authors, we couldn't find
anything that made this clear: the closest I could find was the fact
that TLVs are described in the Discovery Stage of the document (section
5) but not mentioned in the Session Stage (section 6); further there
isn't any way that one could determine if the first field after the
PPPoE header is a TLV type or a PPP protocol identifier.

	 

	I've gone through the list archives and realize that the
original document wasn't warmly received.  I'd like to avoid revisiting
those arguments and simply raise the question as to whether TLVs can
appear in PPPoE Session packets.  If not: what would be a reasonable
alternative: using TLV type codes which correspond to the reserved PPP
protocol values (as per RFC-1661: odd values between 0x0003 and 0x001F
along with 0x00FF)?  I've suggested appending them to PADQ messages
instead since that's already a message type that in their document does
not need to be acknowledged but they want to piggyback the TLVs into
packets in flight rather than deal with separate packets.

	 

	TIA.

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