Re: [CCAMP] Vendor-Specific Application Code in draft-ietf-ccamp-rwa-wson-encode

[Dieter Beller <[e-mail]>]

+1 (support for Option 2)


Thanks,
Dieter

On 29.01.2015 04:46, Fatai Zhang wrote:

Hi Young and all,

 

As an individual, I agree with Adrain, which means I would support Option 2.

 

I think we just need add some sentences to fit Option 2.

 

When s=0 and OI=1, the first 32 (or 48) bits of the Vendor-Specific Application Code field contain an Enterprise Number (or OUI) that defines the context in which the remainder of that field is interpreted. In addition, we need to add some text to describe the process rules that Adrian hinted below.

 

My question is:

Which number (either an Enterprise number or OUI) should be used? I think we just need pick one of them. If an Enterprise Number is used, it needs 32 bits, otherwise 48 bits should be used. If 48 bits are used, it needs to put one line more (32 bits) to the current format of Optical Interface Class.

 

To Young, you can also take a look at [draft-ietf-pce-rfc7150bis], which has some informaiton releated to Enterprise Number.

 

 

 

 

 

Best Regards

 

Fatai

 

 

-----Original Message-----
From: Adrian Farrel [mailto:[e-mail]]
Sent: Saturday, January 24, 2015 2:46 AM
To: Leeyoung; 'Giovanni Martinelli (giomarti)'
Cc: [e-mail]; [e-mail]; [e-mail]
Subject: RE: Vendor-Specific Application Code in draft-ietf-ccamp-rwa-wson-encode

 

> Thanks Adrian for taking this on.

 

Well, we'll get there eventually!

 

> My preference is Option 1.

 

OK.

It is my least favorite option because it is most likely to hit interoperability

issues.

So, if the WG wants to go this way, we will have to work a little to explain how

it works and why it isn't a problem (specifically for the IESG).

 

> It would be hard to catch moving target around this

> area if we were to take Option 2.

 

I see no moving targets at all.

This is how all other vendor-specific fields are handled in protocols.

The way it works is that anyone receiving such a field looks at the first 32

bits and interprets it as an Enterprise Code from the IANA registry (hint: easy

to get and many well-known vendors have them).

If it is an Enterprise Code they recognise they process according to their own

vendor-specific knowledge about the contents.

If they don't recognise the Enterprise Code, they fail the processing.

An Enterprise will often (always?) define their own structure starting with a

version number or something similar, and often using TLVs.

 

What have I missed about moving targets?

 

Cheers,

Adrian

 

_______________________________________________
CCAMP mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/ccamp" rel="nofollow">https://www.ietf.org/mailman/listinfo/ccamp