Re: [addr-select-dt] draft-ietf-6man-rfc3484-revise-02

[Arifumi Matsumoto <[e-mail]>]

Hiromi-san,
thank you for your comment.

First of all, you are commenting about the following rule, right ?

>>> 
>>> I found this sentence a little clumsy:
>>> "When a destination address DA, DB, and the source address of DA
>>>  Source(DA) are on the same subnet and Source(DA) == Source(DB), DNS
>>>  round robin load-balancing cannot function. "
>>> Maybe it can be written a little more clearly?
>> 
>> Sorry, writing clear english is always an assignment for me. X(
>> 
>> How about the following text ?
>> 
>> "When a destination address DA, DB and the source address S are on the
>> same subnet, the present longest matching rule always choose the same
>> destination address."



> Hi list,
> 
> I did only read original Arifumi's document.
> 
> I also feel DA,DB and S description is a little bit confusing.
> My understanding is here;
> 
>             DA(s,d)         DB(s',d')
> 
>              |              |
> subnet P(S,D) -------------------------
> 
> prefix P is same for DA,DB and it can be described as s == s'. Is this
> true?

I failed to see what DA(s,d) means here.
Also the meaning of P(S,D).

> If yes, how about write their network factor into small letter?

I appreciate if you can elaborate on this.

> But as for DNS round robin, current major OSs follows RFC3484 Rule9 and
> neglect DNS information. RFC3483 rule9 is first and if someone wants to
> utilize DNS round robin, 3484 policy table(disable subnet
> prioritization) can be changed is the better indication I suppose.

What you mean here is that you want a switch to enable/disable rule 9, 
that is the longest matching rule, and want to switch on/off it remotely, right ?

My understanding about this issue of rule 9 is,
- This is basically a remote side policy.
  That is, it should be a server side policy, whether rule 9
  is preferable or not.
- The local network administrator basically does not care about this
  behavior of destination address selection, as far as the destination
  site is not in his administration site. 

So, I suppose there is not so much need of local switch of rule 9.
What we need to discuss is what the universal rule about longest
match should be.

> Informative Referece; http://support.microsoft.com/kb/968920
> 
> Don't you put some note above that?

Thank you for information.
I do not quite understand what this configuration does.
If if I put this in my registry, my machine disables rule 9 ?
I do not think such a thing is fully described by the word
"subnet prioritization" in this page.

Regards,

> 
> (2011/03/17 11:31), Arifumi Matsumoto wrote:
>> Tim,
>> 
>> On 2011/03/11, at 9:16, Tim Chown wrote:
>> 
>>> Hi,
>>> 
>>> I have edited the txt file, largely to tidy up the English in a few places.   There's nothing I disagree with in the text.   So you could diff the file with yours to see if you want to apply the changes.
>> 
>> Thank you for your review and comments.
>> I'm sorry that I could not addressed all issues in -02.
>> 
>>> 
>>> Some other comments:
>>> 
>>> We don't cite draft-ietf-6man-addr-select-sol-03 or draft-ietf-v6ops-addr-select-req-07 - would this be worth doing?    Should probably have a final pass through those to ensure everything is captured in the 3484 bis text?
>> 
>> The rfc3484-revise does not mention about the whole problem related to address selection, but it does focus on the present issues for RFC 3484 itself.
>> 
>> It may be helpful if we mention about the whole problem for the readers to grasp the big picture, too. I'll put some texts about them in the introduction, where problem statement and requirement, and the solution documents will be referenced.
>> 
>>> Note the address selection considerations reference is now a WG item: draft-ietf-6man-addr-select-considerations-02; this needs updating in your text.
>> 
>> Applied.
>> 
>>> 
>>> I found this sentence a little clumsy:
>>> "When a destination address DA, DB, and the source address of DA
>>>  Source(DA) are on the same subnet and Source(DA) == Source(DB), DNS
>>>  round robin load-balancing cannot function. "
>>> Maybe it can be written a little more clearly?
>> 
>> Sorry, writing clear english is always an assignment for me. X(
>> 
>> How about the following text ?
>> 
>> "When a destination address DA, DB and the source address S are on the
>> same subnet, the present longest matching rule always choose the same
>> destination address."
>> 
>>> Quite a few of the citations (reference labels) appear after the full stops in sentences, not before.   Corrected in the attached text.
>> 
>> I hope all were corrected.
>> 
>>> 
>>> Do we want to add a rule for 2001:db8:: ?     If 3ffe:: is in there...
>> 
>> I do not know.
>> Do we see 2001:db8:: addresses on a real packets ?
>> 
>>> And I would still love to see a way to add a rule to put privacy address preference in there.   This would be a great way to help answer the concerns in long email threads currently running in other lists :)
>> 
>> Do you think it should be addressed in RFC 3484 revision, or in the policy distribution draft, or totally different way ?
>> 
>> Kindest regards,
>> 
>>> 
>>> Tim
>>> 
>>> <draft-ietf-6man-rfc3484-revise-02-tc.txt>
>>> 
>>> On 9 Mar 2011, at 10:38, Arifumi Matsumoto wrote:
>>> 
>>>> All,
>>>> 
>>>> here attached the -02 version of RFC 3484 revision work.
>>>> I'm very sorry about the long suspension on this item,
>>>> despite your commitment on it.
>>>> 
>>>> If you can comment on it, I think I can incorporate it
>>>> before the cut-off.
>>>> 
>>>> Kindest regards,
>>>> 
>>>> <draft-ietf-6man-rfc3484-revise-02.txt>
>>>> --
>>>> Arifumi Matsumoto
>>>> NGN System Architecture Project
>>>> NTT Service Integration Laboratories
>>>> E-mail: [e-mail]
>>>> TEL [phone] FAX [phone]
>>>> 
>>>> _______________________________________________
>>>> addr-select-dt mailing list
>>>> [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/addr-select-dt
>>> 
>>> _______________________________________________
>>> addr-select-dt mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/addr-select-dt
>> 
>> _______________________________________________
>> addr-select-dt mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/addr-select-dt
> 
> 
> -- 
> ----------
> Ruri Hiromi
> INTEC Systems Institute, Inc.
> _______________________________________________
> addr-select-dt mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/addr-select-dt


--
松本 存史 <[e-mail]>
 NTTサービスインテグレーション基盤研究所
 次世代ネットワーク方式SEプロジェクト
  ネットワークアーキテクチャグループ
  TEL [phone]  FAX [phone]