Keith Moore wrote: > > I was suggesting that SL is an indication that a filtering > policy has > > been applied to this network. > > seems like a *huge* stretch - several of the ideas for using > SL have nothing to do with filtering. also, SL strikes me as > an extremely poor mechanism for communicating filtering policy.
This is demonstrably untrue. SL is not routable between autonomous administrations without explicit coordination to remove ambiguity. With coordination, the regions are no longer completely autonomous, and since there is no ambiguity, applications have no problems using them. Your concern about applications having to work across SL boundaries is fundamentally flawed, because routing will not support it without nat. This is not an argument for nat, and again the only way to avoid nat is to allow simultaneous use of multiple scopes. > > in general an application must assume that there is a > filtering policy may be in place regardless of whether or not > it sees SL addresses. after all, filtering that can affect an > application can exist anywhere in the network, and SL could > at best provide a crude indication of > filtering on the local network. > > the way you determine whether filtering is imposed is by trying to > communicate and having that attempt fail due to an ICMP > 'communication prohibited' message. > > so no, I don't think this flies at all. This perspective assumes all control/feedback is based on outbound traffic. Think 'inbound traffic'. Tony -------------------------------------------------------------------- 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] --------------------------------------------------------------------