Hello, and sorry for the delayed response...
>>>>> On Wed, 19 Sep 2001 12:58:41 -0700,
>>>>> "Richard Draves" <[EMAIL PROTECTED]> said:
>> Our current implementation checks the duplication at the
>> generation of actual addresses as well as at the generation
>> of interface identifiers.
> In my implementation, I don't do any checks directly on the generated
> interface identifier. Instead when I create a temporary address, I check
> for known anycast addresses, collisions with existing addresses, etc. If
> there is a problem, then I generate a new interface identifier and
> repeat. This avoids any timing windows.
> This is not quite equivalent to what the draft specifies. For example
> suppose an interface A has an EUI-64 based interface identifier IDA and
> derived link-local address LIDA. Now interface B is using temporary
> addresses with global prefix P and happens to generate interface
> identifier IDA. Then my implementation will proceed to create a
> temporary address PIDA whereas the draft would prevent interface B from
> using IDA.
> Is there some reason that the more restrictive behavior specified in the
> draft is necessary?
Actually, I don't see any strong reason for the restrictive behavior
(and that's why I asked this question). I implemented the spec this
way just for conformance. So, if there is really no strong reason, I
guess the spec can be loosened.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
