Hi Keith & Brian, thanks for your answers. It seems like there is still some confusion here...
Still a few questions: Looking at the documentation I have that Link Local address is FE80::/64 +64 bits of interface ID. There is no place for the zone. Do you mean that the zone will be assigned by the IfIndex of the interface that the message was recieved by? Another question: MAC address is required to be unique (48 bits) a part of the interface ID. Thus the Link Local assigned to it is unique... So why we bother doing DAD? And also If the above is correct we do not have a problem really... Basically, based on the assumtion that Link Local should be unique everywhere. The only possibility of duplicate Link Local Address is miss-configuration by a human. Can we follow what we are doing in IPv4 in case of wrongly configured address? Basically requiring the network manager to deal with that. Thanks, Shuki -----Original Message----- From: Keith Moore [mailto:moore@;cs.utk.edu] Sent: Thursday, October 17, 2002 8:49 PM To: Brian Zill Cc: sasson, shuki; [EMAIL PROTECTED] Subject: Re: Link Local Address usage for multi-home host. > Yes, this can be an issue, but it's not really a serious problem. forcing apps to deal with scopes at all is a serious problem. being able to compare scopes from different addresses only addresses one small aspect of that problem. though making an app 'protocol independent' is one example of having apps deal with scopes since presumably the app may have to deal with a mixture of v4 and v6 addresses, a mixture of v4-only and v6-only and dual-stack hosts, and various kinds of connectivity (no v4, local v4, global v4) X (no v6, local v6 only, global v6) it should not be assumed that it's appropriate to expect every app to be protocol independent. there will be v4-only apps. there will be v6-only apps. there will be apps that can run one or the other but not both. this is a natural consequence of v6 deployment scenarios. app vendors shouldn't expect that limiting themselves to scoped addresses provides anything in the way of security other than delusion. it's not appropriate for the vendor to assume that the site or local environment is secure. bottom line: there is insufficient justification for expecting apps to deal with scoped addresses. Keith -------------------------------------------------------------------- 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] --------------------------------------------------------------------
