> Proposed test: > > Rule 8: Use longest matching prefix. > If CommonPrefixLen(SA, D) > CommonPrefixLen(SB, D), then choose SA. > Similarly, if CommonPrefixLen(SB, D) > CommonPrefixLen(SA, D), then > choose SB. > > Rule 8 may be superseded if the implementation has other means of > choosing among source addresses. For example, if the implementation > somehow knows which source address will result in the "best" > communications performance. > > Also, an implementation that has knowledge of the prefix lengths > associated to the candidate source addresses MAY choose to > limit longest prefix match to those particular prefix lengths instead > of doing it on the full 128 bits.
I think the above description is insufficient for an implementor to understand what the implementor might optionally do. Thus I'm concerned that folks will read this and do something quite different than we intend. > Rule 9: optional tie breaker > > If the above rules failed to choose a source address, an implementation > MAY > either decide to pick a candidate source address randomly or to take the > smallest one in the lexicographic order. This rule is optional. I don't understand the problem this is trying to solve. I thought there was an argument (and perhaps not agreement) that predictable behavior (having a single host? all hosts with the same configuration?) pick the same source address for a given destination. The above doesn't solve that problem. So what problem is it trying to solve? Confused, Erik -------------------------------------------------------------------- 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] --------------------------------------------------------------------
