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

[Acee Lindem <[e-mail]>]

Manral, Vishwas wrote:
> Hi Acee,
>
> Though you can have vendor specific TLV encodings, the issue is there may be
> a clash in the TLV or Sub-TLV values.
>
> By using an SMI enterprise code or OUI to distinguish between various vendor
> specific implementations, we have one value that can be used by vendors
> instead of taking any TLV values, and we do not confuse information by
> various vendors, which are used for different purposes.

Hi Vishwas,

I understood the motivation. I think the SMI enterprise code is a better
choice than the OUI since it is administered by IANA. However, I'm against
the basic premise of allowing every vendor to roll their own TE LSAs and
carry them in OSPF. I believe doing so could impact 1) Interoperability (there
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.

Thanks,
Acee

>
> Thanks,
> Vishwas
>
> -----Original Message-----
> From: Acee Lindem [mailto:[e-mail]]
> Sent: Wednesday, May 28, 2003 10:07 AM
> To: [e-mail]
> Subject: Re: Vendor attributes in TE LSAs - (Sent first reply
> prematurely)
>
>
> References:
> <[e-mail]>
> <[e-mail]>
> Content-Type: text/plain; charset=us-ascii; format=flowed
> Content-Transfer-Encoding: 7bit
>
> Udo,
>
> Aside from any technical details, I don't agree with this draft
> philosophically. In essence, it is a proposal to standardise on
> a mechanism to be non-standard. More specifically, it removes
> the OSPF WG, TE WG, and IETF in general from reviewing the vendor
> specific TLVs. The contents, size, and refresh rate of these TLVs
> are unknown.
>
> By definition, there is no interoperability. The same set of problems
> will undoubtedly be solved in multiple ways by different vendors using
> different vendor specific TLV encodings.
>
> Acee Lindem wrote:
>
>>Udo,
>>
>>I don't support this enhancement since it essentially removes any
>>
>>
>>Udo Neustadter wrote:
>>
>>
>>>Hi all,
>>>
>>>I am working on a GMPLS implementation, and part of my problem is the
>>>addition of company specific data to the TE LSAs. The Internet-Draft
>>>http://www.ietf.org/internet-drafts/draft-udo-ospf-vendatt-00.txt
>>>proposes an interoperable way to solve the following two issues:
>>>  1. Companies that already applied with IANA for an SMI Network
>>>management enterprise code do not need to re-apply for sub-TLV values
>>>from the pool of numbers reserved for private use
>>>  2. Allows private attributes/data to be embedded in the TE router LSA
>>>(the one TE LSA that contains the router address TLV).
>>>
>>>I would like for the draft to be considered part of this working group.
>>>This work is an extension to the work done in
>>>http://www.ietf.org/internet-drafts/draft-katz-yeung-ospf-traffic-09.txt
>>>and
>>>http://www.ietf.org/internet-drafts/draft-ietf-ccamp-ospf-gmpls-extensio
>>>ns-09.txt.
>>>
>>>Thanks in advance for your support.
>>>
>>>Udo
>>>
>>
>>
>>--
>>Acee
>>
>
>
>
> --
> Acee
>


--
Acee