Re: [trill] Suresh Krishnan's Discuss on draft-ietf-trill-irb-13: (with DISCUSS and COMMENT)

[Donald Eastlake <[e-mail]>]

Hi Suresh,

On Thu, Jun 30, 2016 at 11:14 PM, Suresh Krishnan
<[e-mail]> wrote:
> Hi Donald,
>    Your proposed changes and clarifications look good to me. I will look over
> the new version when it is posted.

Version -14 has been posted with the intent of resolving your DISCUSS.

Thanks,
Donald
===============================
 Donald E. Eastlake 3rd   [phone] (cell)
 155 Beaver Street, Milford, MA 01757 USA
 [e-mail]

> Thanks
> Suresh
>
> On 06/30/2016 02:13 PM, Donald Eastlake wrote:
>> Hi Suresh,
>>
>> Thanks for your comments. See below.
>>
>> On Wed, Jun 29, 2016 at 7:08 PM, Suresh Krishnan
>> <[e-mail]> wrote:
>>> Suresh Krishnan has entered the following ballot position for
>>> draft-ietf-trill-irb-13: Discuss
>>>
>>> ...
>>>
>>> ----------------------------------------------------------------------
>>> DISCUSS:
>>> ----------------------------------------------------------------------
>>>
>>> * Section 6 has a few errors that need to get fixed before this document
>>> goes forward. e.g. It is not clear what a "192.0.2.0/32" subnet means
>>> especially since the only host shown to be on the subnet 192.0.2.2 cannot
>>> obviously fall inside the subnet range. The /32 needs to be replaced with
>>> something shorter depending on what the authors/WG intended (say a /24).
>>
>> Yes. That should have been "/24".
>>
>>> * RB2 seems to be advertising ES2s IPv4 address 198.51.100.2/32 instead
>>> of the prefix of the subnet while RB1 seems to be advertising the the
>>> IPv4 prefix of the ES1 subnet. One of these is wrong. Not sure which one
>>> is intended.
>>
>> It should be the prefix in both cases.
>>
>>> * What is the rationale for using a /112 IPv6 prefix for numbering an
>>> IPv6 link with hosts? Things like SLAAC (RFC4862) will not work in such
>>> links. Is there a reason the authors want to use a longer than /64?
>>> Please read RFC7421 for advantages of using a /64 instead and to find out
>>> what things break if you do not use a /64.
>>
>> The Distributed Layer 3 gateway specified in this draft is expect to
>> primarily be used in data centers where I would expect everything to
>> be strictly configured by an orchestration system. Thus stateless
>> autoconfiguration seems less likely and I suspect it just wasn't
>> through of. However, I don't see a problem with changing this to "/64"
>> and the mechanism specified could be used in other contexts outside
>> data centers.
>>
>>> ----------------------------------------------------------------------
>>> COMMENT:
>>> ----------------------------------------------------------------------
>>>
>>> Section 5: What does "Layer 2 routing" mean in this context?
>>
>> Previously standardized TRILL routing, which is based on destination
>> MAC address as mapped into a TRILL nickname. The wording can be
>> clarified.
>>
>>> Sections 7.3 & 7.4: What is the point of including these sub-TLVs if no
>>> prefix is being advertised? (The Total Length=0 case specified in the
>>> document)
>>
>> Sometimes a zero length has some special meaning. The draft just says
>> that it has the obvious meaning although it is not a particularly
>> useful value.
>>
>> Thanks,
>> Donald
>> ===============================
>>   Donald E. Eastlake 3rd   [phone] (cell)
>>   155 Beaver Street, Milford, MA 01757 USA
>>   [e-mail]