> All Alice knows is that Bob is not asking for the same protection that > Alice is. There are too many semantics being assumed about the magic > bit. This specific response assumes that there is an implicit > bi-directional requirement that if A sends the bit set to B, then B must > use the same bit set in its address back to A. Yet in other messages all > the magic bit says it that A doesn't want B to interpret any BU messages > that can't be validated. Which is it??? Certainly there is no > requirement for B to tell A that it shouldn't interpret a BU from it, > since B may have a different policy at the moment.
To my understanding it is becouse Alice uses the "strong" exchange, indicated in IID that doesn't want Bob to perform a BU with a "weak" exchange. Bob policy is independent of this discussion, becouse Bob MUST NOT perform a BU under those circunstances. > If you really want a protocol then create one, don't assume that a > single bit sent in one direction can carry the semantics of a > bi-directional protocol. In particular for the problem being solved, > don't ever assume that the policies have to be symmetric. All this > mechanism is trying to do is let the CN know how to interpret a BU sent > on behalf of the MN should be interpreted in a specific way. That is a > unidirectional statement, but might include a feedback path to let the > MN know if the message was received and the CN will honor it. The > opposing unidirectional statement MUST NOT be required to have the same > interpretation. A feedback path is available using the status field of the Binding. So if Bob has not resources or intentions to follow Alice wishes there is a mechanism to provide feedback. /aep -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------