[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: RE: [NSIS] QSPEC extensibility and re-use of multiple instances of defined parameters



Hi, Cornelia:

Whatever you may say, the draft is NOT technically consistent with respect to QSPEC extensions specific to QOS models as you yourself has admitted (e.g., you tried .. ".. accommodate as many QoS Models as possible .."

You will be better off to delete those QOS model-specific specific parts of draft as John L. has also admitted that it will create problems if domains are NOT constructed carefully.
 
Tseno has suggested to generalize the extensions related QOS models for making draft consistent. Jerry suggested to use multiple instances to take care of this problem. John L. has gone further in explaining why Jerry’s suggestions for multiple instances of the same parameter is NOT acceptable.
 
So, where is the consensus unless you remove those technical inconsistencies through deleting the QOS specific extension part of the draft?
 
Best regards,
Radhika
 
Title: Message

Hi Radhika,

 

frankly, I dont see the point of your questions, and my impression is nor does anybody else. Rather than taking a lot of time trying to explain (again) everything I'd like to quote rough working group consensus that the QSPEC is ok the way it is.

 

Cornelia

-----Original Message-----
From: Roy, Radhika (AEAD) [mailto:RoyR at saic-abingdon.com]
Sent: Mittwoch, 17. August 2005 20:17
To: Kappler, Cornelia
Cc: Tseno Tsenov; nsis at ietf.org
Subject: RE: [NSIS] QSPEC extensibility and re-use of multiple instances of defined parameters

Cornelia:

In response to one of Tseno's questions, you have mentioned as follows:

".. accommodate as many QoS Models as possible .."

In fact, it is NOT true. Even the QOS model like Y.1541 that has been trying to be accommodated fully in the draft has been failed. Please see the following draft:

http://www.ietf.org/internet-drafts/draft-ietf-nsis-y1541-qosm-00.txt

The above draft shows that more extensions in QSPEC template are needed.

The basic question is this: Why do you want to include the QOS model-specific extensions in the "generalized" QSPEC template draft?

What you have done is this: Included Y1541 QOS model (which is also incomplete) and DSTE QOS model (a cheap shot - basically only one parameter).

Tseno has mentioned the most fundamental QOS models of the IETF: IntServ and DiffServ. You have failed to explain why you have not included those in the draft.

What do you mean that you have accommodated as many QOS models as possible?

Why do you NOT "delete" all so-called QOS model-specific extensions from this draft?

The way you should proceed is as follows:

1.      Make the draft-ietf-nsis-y1541-qosm-00.txt complete including "all" extensions QSPEC specific to y1541using all "specific values." Also, show how to create y1541 domains using specific values without contradicting the end-to-end values from end users' and/or applications' point of view so that people can implement without creating any confusions or contradictions.

2.      Do the same writing another draft for DSTE model as explained in item 1.

3.      Delete all QOS model-specific extensions in the generalized QSPEC draft.

In this way, all drafts will be clean.

If you have any questions, please let us know.

Best regards,

Radhika

PS: I also explained in earlier emails (John L. also agreed) that the use of both generalized QSPEC template and QOS model-specific QSPEC template will create contradictions if domains are not created carefully. The present QSPEC draft does not explain how to create respective QOS domains using specific values in the first place.

-----Original Message-----
From: nsis-bounces at ietf.org [mailto:nsis-bounces at ietf.org] On Behalf Of Kappler, Cornelia
Sent: Wednesday, August 17, 2005 4:22 AM
To: Tseno Tsenov; Ash, Gerald R (Jerry), ALABS; nsis at ietf.org
Subject: RE: [NSIS] QSPEC extensibility and re-use of multiple instances
of defined parameters

Hi Tseno - and John and Jerry,

so there doesnt seem to be much concrete use for this additional kind of

extensibility. Therefore my suggestion is to not support it.

Regarding your (Tseno) 2nd paragraph:

In the QSPEC we tried to balance two mutually exclusive requirements

- accommodate as many QoS Models as possible

- make the QSPEC as widely understood as possible without defining too

many parameters.

We solved this problem by defining mandatory parameters which we expect

to be used in most QOSMs (e.g. bandwidth); and by requesting all QNEs to

_understand_ the mandatory parameters. That is, for a QOSM that has no

use of the mandatory parameters (which is quite unlikely by the way)

these parameters _dont need to be included_ in the QSPEC. However this

makes it unlikely that this QSPEC will be understood outside the domain

of this QOSM.

I guess we have to explain this better in our document as this is not

the first time somebody (particularly someone really familiar with NSIS)

didn't really get the idea.

Cornelia

 

_______________________________________________
nsis mailing list
nsis at ietf.org
https://www1.ietf.org/mailman/listinfo/nsis