Measuring impairments [Was: Updated Draft Liaiosn to Q6/15]

"Adrian Farrel" <adrian@olddog.co.uk> Tue, 10 March 2009 16:48 UTC

Envelope-to: ccamp-data0@psg.com
Delivery-date: Tue, 10 Mar 2009 16:49:04 +0000
Message-ID: <F121CD66F6334664B3BCE0DF21CAB4A6@your029b8cecfe>
Reply-To: "Adrian Farrel" <adrian@olddog.co.uk>
From: "Adrian Farrel" <adrian@olddog.co.uk>
To: "Giovanni Martinelli" <giomarti@cisco.com>, "Malcolm Betts" <betts01@nortel.com>, "O'Connor, Don" <don.oconnor@us.fujitsu.com>
Cc: <ccamp@ops.ietf.org>
Subject: Measuring impairments [Was: Updated Draft Liaiosn to Q6/15]
Date: Tue, 10 Mar 2009 16:48:25 -0000
MIME-Version: 1.0
Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=response
Content-Transfer-Encoding: 7bit

I'm going to try to answer all of the comments about measuring impairments 
in one email.

I'm arguing all of this from an abstract point of view. I want to "out" in 
advance of the meeting as much of opinion held in CCAMP. I do not believe it 
is valuable to go into the meeting expressing what we think may be Q6's 
view. Instead, we need to say what it is people in CCAMP may want to do. 
Then we can get Q6 feedback on whether that is practical and what the 
concerns are.

So...

The ability to measure optical impairments on an active path is claimed by 
several vendors. I am not in a position to judge whether they are successful 
or not.

Giovanni reasonably asks "what exactly you mean by *ability to measure*?"

We are proposing protocol extensions that allow nodes to distribute 
information about optical impairments. It is not our business to define from 
where this information is gathered. We can observe that the information 
might be configured, might be measured during network provisioning and held 
static, might be determined by a node applying some algorithm to configured 
on pre-measured information, or might be measured dynamically. So we can 
choose between:
- optical impairments can be advertised, but cannot be updated
- optical impairments can be advertised, and can be updated

If we choose the first of these, it seems that we are shutting out what some 
people want to be able to do. If we choose the latter, we are not requiring 
anyone to update the information they advertise, but we are allowing this to 
be done if a node chooses to do so.

To answer Don specifically, I see no proposal in CCAMP about which 
impairments could be measured or how they would be measured. But, to turn 
this point around, I do not believe that CCAMP should say "you must not 
measure an impairment". As Don says, this is outside our remit.

 Malcolm's suggestion doesn't cut it for me.
By saying "We understand that Q6 currently has no requirement to measure 
impairments after the transport equipment is deployed" we miss the point. 
The point is not what Q6 requires or does not require, but is what CCAMP 
requires.

So I wonder what is wrong with the statement (in the context of describing 
what CCAMP wants to do) that "There is no requirement to measure 
impairments."

Don objected specifically to...
>    However, if an implementer chooses to measure impairments
>    on their device, this should not be prohibited, and should be
>    accommodated.

How would it be if we defered the practicality of such measurements to the 
ITU? We could then write...

    However, if an implementer chooses to measure impairments
    on their device, and this can be achieved within the mechanisms
    and definitions defined by the ITU-T, then this should not be
    prohibited by the CCAMP protocol mechanisms, and should be
    accommodated within GMPLS.

Cheers,
Adrian