Re: [PCN] Terminology

[Michael Menth <[e-mail]>]

Hi Phil and others,

I have the feeling that everybody can live with the term "admissible rate".

Regarding the terms "sustainable or supportable rate" I suspect that 
some people are not happy about them because they implicate some "right" 
to always use that bandwidth although this should rather be an 
exception. I can understand that. The term "configured-termination-rate" 
anticipates the action how the traffic rate is reduced when the 
respective rate is exceeded although options exist, therefore, this 
expression isn't good either. A neutral solution to that problem is 
"tolerable rate" which avoids all the mentioned misleading connotations. 
Any objections?

Regards,

    Michael


[e-mail] wrote:
> Hi
>
> It seems more intuitive to me that the terminology should come as "sets"
> of very similar terms that relate to the same mechanism. IE related
> terms should obviously be related without having to think about it:-
>
> admission control | configured-admissible-rate | admission-marking (&
> admission-marked-rate)
>
> termination control | configured-termination-rate | termination-marking
>
> I find it not instantly obvious that your suggested terms are related:
> {termination control | excess-traffic marking | supportable-rate} 
>
>
> A few more comments in-line
>
> Thanks
> phil
>
>
>   
>> -----Original Message-----
>> From: Michael Menth [mailto:[e-mail]]
>> Sent: 10 July 2007 07:44
>> To: [e-mail]
>> Subject: [PCN] Terminology
>>
>> Hi,
>>
>> I would like to discuss (again) some terminology for PCN. To get
>>     
> things
>   
>> done quickly, we stuck in the architecture draft to the working
>> nomenclature from the CL draft and left the exact names for further
>> discussion. See Section 2 of
>>
>>     
> http://www.ietf.org/internet-drafts/draft-eardley-pcn-architecture-00.tx
> t
>   
>> We proposed an alternative nomenclature in Section 1.3 of
>> http://www.ietf.org/internet-drafts/draft-babiarz-pcn-3sm-00.txt
>>
>> The main differences are the following four names, all other terms are
>> fine:
>>
>> "admissible rate" instead of "configured-admission-rate":
>>     
>
> [phil] the draft actually uses configured-admissible-rate. Dropping
> 'configured' would be ok with me. 
>
>   
>> It's the rate that CAN be admitted. Then it is self-understood that
>>     
> the
>   
>> rate is not measured but configured.
>>
>> "supportable rate" instead of "configured-termination-rate":
>> It's the rate that CAN be supported. Then it is self-understood that
>>     
> the
>   
>> rate is not measured but configured. Furthermore, an alternative
>>     
> action
>   
>> for flow termination is rate reduction for which the name
>>     
> "termination"
>   
>> is not appropriate.
>>     
>
> [phil] strictly it isn't the rate that can be supported; traffic can be
> supported up to the rate configured for the DS class used by the PCN
> traffic. It's the rate above which you start termination-marking pkts. 
>
> Perhaps we need something like terminatable-rate or
> terminatable-marking?! (I will be back)
>
> Yes, the charter says that after re-chartering the WG "may also consider
> to investigate additional response mechanisms that act on
> (pre-)congestion information". So your same point could be made that
> admission-marking is an inaccurate term, because in the future the WG
> might consider an action other than admission control. However,
> personally I don't think potential future charter should influence
> terminology.
>
>   
>> "admission-stop marking" instead of "admission-marking":
>>     
>
> [phil] I don't object too much to admission-stop-marking (or
> admission-control-marking) instead of admission-marking. Personally I
> find the latter a bit simpler.
>
>   
>> The meaning of the marking is that admission should be stopped. The
>>     
> name
>   
>> "admission-marking" rather suggests the opposite. 
>>     
>
> [phil] isn't it self-understood that pkts are marked at higher traffic
> rates and not marked at low traffic rates?
>
>   
>> This is also in line
>> with the single-marking approach.
>>
>> "excess-traffic marking" instead of "termination-marking":
>> The name "termination-marking" is fine as long as flow termination is
>> the only option to reduce the rate of ingress-egress-aggregates when
>>     
> the
>   
>> supportable rate of a link is exceeded. However, an alternative
>>     
> reaction
>   
>> to this marking is flow rate reduction (although not in this charter).
>> Therefore, the name of that marking should be neutral with
>> regard to the action.
>>
>> Regards,
>>
>>     Michael
>>
>> --
>> Dr. Michael Menth, Assistant Professor
>> University of Wuerzburg, Institute of Computer Science
>> Am Hubland, D-97074 Wuerzburg, Germany, room B206
>> phone: [phone], fax: [phone]
>> mailto:[e-mail]
>> http://www3.informatik.uni-wuerzburg.de/research/ngn
>>
>>
>>
>> _______________________________________________
>> PCN mailing list
>> [e-mail]
>> https://www1.ietf.org/mailman/listinfo/pcn
>>     

-- 
Dr. Michael Menth, Assistant Professor
University of Wuerzburg, Institute of Computer Science
Am Hubland, D-97074 Wuerzburg, Germany, room B206
phone: [phone], fax: [phone]
mailto:[e-mail]
http://www3.informatik.uni-wuerzburg.de/research/ngn



_______________________________________________
PCN mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/pcn