Re: [trill] seeking help to understand a line from TRILL-ESADI draft

[Radia Perlman <[e-mail]>]

Oh right!  You just learn of your ESADI neighbors based on the core
instance LSPs.  Sorry about the confusion.

Radia

On Wed, Jul 18, 2012 at 12:23 PM, Donald Eastlake <[e-mail]> wrote:
> Currently there are no Hellos in ESADI. Participation is determined
> from the core IS-IS link state database.
>
> Thanks,
> Donald
> =============================
>  Donald E. Eastlake 3rd   [phone] (cell)
>  155 Beaver Street, Milford, MA 01757 USA
>  [e-mail]
>
>
> On Wed, Jul 18, 2012 at 3:12 PM, Radia Perlman <[e-mail]> wrote:
>> Actually, perhaps ESADI shouldn't bother with CSNPs at all.  If each
>> RB participating in the ESADI instance included the sequence number of
>> its own LSP in its Hello for that ESADI instance, wouldn't that work?
>> We need to have each RB (in the ESADI instance) send Hellos
>> periodically anyway.
>>
>> Radia
>>
>> On Wed, Jul 18, 2012 at 11:15 AM, Thomas Narten <[e-mail]> wrote:
>>> Seems to me that the MAY sentence all by itself is not helpful.
>>>
>>> If sending occasionaly ESADI-CSNPs increases robustness, IMO, the spec
>>> should say that and explain which scenarios its useful in. And then,
>>> MAY doens't seem right either. SHOULD would seem to make more sense...
>>>
>>> The RFC 2119 defintion of MAY says:
>>>
>>> 5. MAY   This word, or the adjective "OPTIONAL", mean that an item is
>>>    truly optional.  One vendor may choose to include the item because a
>>>    particular marketplace requires it or because the vendor feels that
>>>    it enhances the product while another vendor may omit the same item.
>>>    An implementation which does not include a particular option MUST be
>>>    prepared to interoperate with another implementation which does
>>>    include the option, though perhaps with reduced functionality. In the
>>>    same vein an implementation which does include a particular option
>>>    MUST be prepared to interoperate with another implementation which
>>>    does not include the option (except, of course, for the feature the
>>>    option provides.)
>>>
>>> which doesn't seem to match the explanation for why the MAY sentence
>>> is in the document.
>>>
>>> Thomas
>>>
>>> Donald Eastlake <[e-mail]> writes:
>>>
>>>
>>>> Hi Swet,
>>>
>>>> On Tue, Jul 17, 2012 at 5:59 AM, Swet kumar <[e-mail]> wrote:
>>>> > Hi Authors,
>>>> >
>>>> >
>>>> > I am not able to understand the meaning of the following line in this draft:
>>>> > http://tools.ietf.org/pdf/draft-ietf-trill-esadi-00.pdf
>>>
>>>> Thanks for reviewing this document. It was just recently posted as a
>>>> WG document and I think it does need some more work.
>>>
>>>> > “For robustness, if an ESADI instance has two or more ESADI neighbors
>>>> >
>>>> >    and is not DRB and it receives no ESADI-CSNP PDUs for at least the
>>>> >
>>>> >    CSNP Time (see Section 6.1) of the DRB, it MAY transmit an ESADI-
>>>> >
>>>> >    CSNP.”
>>>
>>>> Do you really no understand what the sentence says? Or is it that you
>>>> don't understand what the reason is for this behavior?
>>>
>>>> The idea is that if the DRB on an ESADI virtual link is being flakey
>>>> or the virtual link is broken then, to maintain LSP synchronization,
>>>> it would be good for some other TRILL switch on that ESADI virtual
>>>> link to occasionally send a CSNP. The multi-hop ESADI pseudo-link is
>>>> inherently somewhat less reliable that a link between adjacent TRILL
>>>> switches. (You might think this could result in a flurry of CSNPs but
>>>> note that, due to the jitter in the IS-IS reliable flooding mechanism,
>>>> if there were several non-DRB TRILL switches on the link, as soon as
>>>> one sends a CSNP it will inhibit the others for a while.)
>>>
>>>> Anyway, although it doesn't make much difference. It should probably
>>>> say "one" instead of "two" in the text and it should not just be that
>>>> it has not received a CSNP but that it has not received or sent a
>>>> CSNP. So, I think the wording should be changes to:
>>>
>>>>    "For robustness, if an ESADI instance has one or more ESADI neighbors
>>>>    and is not DRB and it does not receive or send an ESADI-CSNP PDU
>>>>    for at least the CSNP Time (see Section 6.1) of the DRB, it MAY
>>>>    transmit an ESADI-CSNP."
>>>
>>>> Thanks,
>>>> Donald
>>>> =============================
>>>>  Donald E. Eastlake 3rd   [phone] (cell)
>>>>  155 Beaver Street, Milford, MA 01757 USA
>>>>  [e-mail]
>>>
>>>> > This line is a part of section 4.1 (Sending of ESADI PDUs) on page number
>>>> > 12.
>>>> >
>>>> > Please explain me what this line tries to convey.
>>>> >
>>>> >
>>>> > Regards,
>>>> >
>>>> > Swet
>>>> _______________________________________________
>>>> trill mailing list
>>>> [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/trill
>>>
>>>
>>> _______________________________________________
>>> trill mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/trill
>>>