Re: Evaluation: draft-dasilva-l2tp-relaysvc - L2TP Active Discovery Relay for PPPoE to Proposed Standard

[Thomas Narten <narten@us.ibm.com>]

> 2 Nits, both in section 2.3:  "is simply ensure" should probably
> be "is simply to ensure" and "given that it is only the
> source of the encrypted and data" seems to be missing a noun.

ack.

> I'm also not generally in favor of the kind of overloading that 2.3
> implies you might do with your "arbitrary data" (since it tends
> to get a lot less arbitrary when you do so), but since they only
> said someone might do it and didn't define a mechanism we
> could actually check for corner cases, I don't think there's
> anything we could do about it.

Can you expand? I don't understand the concern.

The basic idea here is that the "abitrary data" is internal state of
the sender. No one else uses that data (or parses it) it is simply
echoed back in the response traffic, allowing the relaying device to
insert state about the "request response pair" in the message and then
extract it from the response. This allows it to put the state in the
message rather than on the device (i.e, devices are more stateless).

Thomas