[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: comments on draft-daniel-ipv6-ra-mo-flags-00.txt



Hi Jinmei,

JINMEI Tatuya / ???? wrote:
>>>>>>On Wed, 11 Aug 2004 19:59:43 +0530, 
>>>>>>Syam Madanapalli <syam@samsung.com> said:
>>>>>
> 
>>>>M/O flags indicate the avaialbility of the respective service, so if
>>>>a router advertises the M/O flags bits ON, I think we should OFF
>>>>them if and only if the same router advertises again to OFF. It is
>>>>administartor problem if one advertises with bits ON, and other
>>>>router with bits OFF.
>>>
>>>How do you propose the host keeps track of which router advertised
>>>which bits?  This kind of tracking is not done at the moment.
>>
> 
>>True, this requires one needs maintain the M/O flags information per router
>>basis. And this leads to implementation complexity. But I think this method
>>provides robust implemention. I am not sure how people are managing if
>>two different routers on same link advertises two different Link MTU. I
>>think the problem could be similar in both the cases.
> 
> 
> (I'm responding to this message after reading all the other messages
> in this thread)
> 
> I'm afraid people may forget that it is administrator's responsibility
> to ensure the consistency among RA parameters from multiple routers
> in the same single link (see Section 5.6 of RFC2462 - while the RFC
> does not explicitly say it's admin's responsibility, but I strongly
> believe it's the background intent).  Section 5.6 of RFC2462 also
> clearly specifies the behavior when a host happens to receive
> inconsistent parameters:
> 
>    If inconsistent information is learned different
>    sources, the most recently obtained values always have precedence
>    over information learned earlier.
> 
> An implementation might explore an implementation dependent trick like
> combining routers and parameters to mitigate the bad effect of the
> admin error.  However, I'd really want to make it outside the scope of
> the M/O document.

I agree with this,

In this case, we need to ensure that the router advertisement which
we're basing the decision on has the flags set.

It may even be worthwhile explicitly stating that management of
flags from multiple routers is not in scope.

Greg




--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------