Re: Vendor attributes in TE LSAs - (Sent first reply prematurely)

[Acee Lindem <[e-mail]>]

Tony Przygienda wrote:
> Acee Lindem wrote:
>
>
>>Tony Przygienda wrote:
>>
>>
>>>Acee Lindem wrote:
>>>
>>>
>>>
>>>>isn't any) 2) Scalability (LSA contents, size, and refresh frequency
>>>>are
>>>>unknown) and 3) Deployment - Debugging a multi-vendor network with many
>>>>different vendor specific LSAs will be difficult.
>>>
>>>
>>>
>>>ok, but then from experience, they will _just_ take the TLVs they need
>>>without
>>>letting you know. Suit yourself ;-)
>>
>>
>>Maybe so - but at least this activity won't be sanctioned by the OSPF WG.
>
>
> this is just a semantic or maybe semiotic split-of-hairs IMHO. OSPF-WG
> or any
> other WG-approval is not a consumer brand and does nothing for the vendor,
> espeically if he's pursuiting a proprietary value-add-on in the protocol.
>
>
>>
>>Perhaps the IANA allocation of TLVs and Sub-TLVs needs to be at the front
>>end of the process rather than the backend.
>
>
> yes, but is easier for vendors to grab an 'experimental' OUI (which is a
> no-time-involved
> process) and once they figured
> out what it does for them and care about interoperability, ask for an
> 'official' sub-TLV.
> It will be a well-understood process by both vendors AND their
> customers. Customers
> will understand that an 'experimental' gives them no guarantee as to
> itneroperability as
> Ran very clearly formulated and has been put down in the
> isis-experimental draft.
> Otherwise they'll ask you for a sub-TLV or probaly not even that (since
> it needs time, so
> they
> just grab any number that looks free)
> and of course not tell you what it does. Difference against just
> semantical IMHO
> and in terms of
> reality the output is the same

I'd be prefer an experimental range of TLVs that are
allocated from IANA for specific experimental purposes rather
than a single TLV giving carte blanche to advertise anything
in OSPF. I'm less cynical that the output will be the same. Hopefully,
if an experimental TE application is commerically viable it
will evolve into a standard.

>
>     thanks
>
>     -- tony
>


--
Acee