RE: [mobile-ip] RE: RFC 2462 DAD optimization

[Dr. Subrata Goswami]

The HA would deny that node use of the link.  This is quite unlikely to ever happen, though.   Why is that ?       I don't understand what you mean here.  Which addresses are useless?   Because you are ignoring the pre-fix (which can be 64 bits).           -----Original Message----- From: Charlie Perkins [mailto:[EMAIL PROTECTED]] Sent: Saturday, June 01, 2002 10:41 PM To: Dr. Subrata Goswami Cc: [EMAIL PROTECTED]; 'IPng Working Group' Subject: Re: [mobile-ip] RE: RFC 2462 DAD optimization     Hello Subrata,   I reckon I am missing your point, so here is a request for clarification:   "Dr. Subrata Goswami" wrote:   > I agree with Dave that, doing a DAD for "link-local" type of post-fix > address is a unnecessary restriction.   The choice seems to be to do DAD once for a "link-local" or "subnet-local" address (even if the node didn't configure one (?!?)), vs. doing DAD for _every_ configured address.   So, the thing you label a "restriction" seems to bring a large measure of freedom, if I understand you correctly.   > 1. There might be global addresss manually configured with the same > post-fix but  with different pre-fix as the link-local address of a > node. What would HA do if a node like that moves into its territory  ?   The HA would deny that node use of the link.  This is quite unlikely to ever happen, though.   > 2. For 64-bit post-fix it would render  2**64 perfectly good addresses > useless.   I don't understand what you mean here.  Which addresses are useless?     Regards, Charlie P.