pppoe tag length
RFC 2516 said pppoe tag should be in the format of (type,lenght,value)
with the length being the length of the value. Would it not be more
beneficial if the length is for the whole tag ? It seems like that
specification is not orthogonal.
Any comment will be appreciated.
Re: pppoe tag length
"sumlut" <firstname.lastname@example.org> writes:[color=blue]
> RFC 2516 said pppoe tag should be in the format of (type,lenght,value)
> with the length being the length of the value. Would it not be more
> beneficial if the length is for the whole tag ? It seems like that
> specification is not orthogonal.[/color]
Yep, it sure would have been nice to have defined it that way.
But, it wasn't. And it's deployed, so that's not fixable.
That's hardly the only problem. A much bigger problem with PPPoE is
that there's just a single session ID number that is assigned by the
server alone. This makes demultiplexing connections very hard if
you're a client that has sessions established with multiple servers.
The way multiplexing protocols are normally designed is to have each
side provide the other with a session ID -- in effect to say, "when
you send data back to me, use this number to identify it." This makes
the implementation a straightforward table lookup on input, followed
by a simple comparison to check the sender and the state.
With PPPoE, you can't do that. If two peers happen to pick the same
ID number for the data they're going to send you, you have to live
with it, and use the sender's Ethernet MAC address to disambiguate the
sessions. It's pretty messy.
PPPoE, for what it's worth, is not the product of an IETF effort. It
was not subjected to substantial review within that group. The
standards-track protocol for tunneling PPP is L2TP.
James Carlson, KISS Network <email@example.com>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677