Re: [RE] Grand master identifier ==>(evolved to) overprovisioning
All,
I received an email note, shortly after previous posting:
>> David -- both URL links provided below do not seem to work.
I'm not sure why the IEEE site was down.
To deal with such possibilities, I have also made this
available on my personal web site, listed below:
http://dvjames.com/esync/dvjRePaper2005May03.pdf
Since I have no storage or BW limits, this should work.
I've also CC'd this to David Law, who perhaps can
provide insight on loss of the IEEE site posting.
For context, the relevant email chain is listed below:
DVJ
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
David V James <dvj@ALUM.MIT.EDU>
Sent by: owner-stds-802-3-re@IEEE.ORG
06/05/2005 02:05 PM
Please respond to David V James
To: STDS-802-3-RE@LISTSERV.IEEE.ORG
cc:
Subject: Re: [RE] Grand master identifier ==>(evolved to)
overprovisioning
Hugh,
In response to the following statement:
>> I expect that you, or someone in the study group,
>> should find the references to the existing standards
>> for provisioning, admission control, policing and QOS;
>> demonstrate why they don't meet your needs
Perhaps you missed a previous email?
For convenience, that message from DVJ is repeated below:
>> -----Original Message-----
>> From: owner-stds-802-3-re@IEEE.ORG
>> [mailto:owner-stds-802-3-re@IEEE.ORG]On Behalf Of David V James
>> Sent: Thursday, May 05, 2005 4:17 PM
>> To: STDS-802-3-RE@LISTSERV.IEEE.ORG
>> Subject: Re: [RE] Grand master identifier ==>(evolved to)
overprovisioning
>>
>> Assuming 75% time-sensitive traffic with arbitrary topology
>> and loading (subject the the aforementioned 75% rule), then
>> I believe proofs are contained within "Annex G" of:
>> http://www.ieee802.org3/re_study/material/index.html
In response to the following statement:
>> and then propose something tangibly different.
Something tangibly different can also be found in:
http://www.ieee802.org3/re_study/material/index.html
But, for this purpose, please exclude Annex G, which
illustrates the need but is not part of the solution.
It meets "tangibly different" in the sence that it meets the
"Assuming 75% time-sensitive traffic with arbitrary topology
and loading (subject the the aforementioned 75% rule)"
criteria and existing standards do no.
Its violates "tangible different" in that the working paper
is evolutionary, not revolutionary, heavily leveraging
concepts and techniques from existing standards. But,
I hope that (in this context) this is a desirable trait.
Cheers,
DVJ
-----Original Message-----
From: Hugh Barrass [mailto:hbarrass@cisco.com]
Sent: Friday, May 06, 2005 5:58 AM
To: David V James
Cc: STDS-802-3-RE@LISTSERV.IEEE.ORG
Subject: Re: [RE] Grand master identifier ==>(evolved to) overprovisioning
David (and others),
Some personal advice from me to anyone who wishes to propose a change to an
existing standard:
The burden of proof should always be on those who want to make the change.
It is necessary that the Study Group work should include serious analysis of
the current standards to demonstrate that there are deficiencies to be
addressed. If there is no one in the study group with sufficient expertise
to explore current standards then I question whether there is anyone with
sufficient expertise to make new standards to replace them. I expect that
you, or someone in the study group, should find the references to the
existing standards for provisioning, admission control, policing and QOS;
demonstrate why they don't meet your needs and then propose something
tangibly different.
In short, I will not be satisfied by the demand, "We don't understand
current standards, so we want to make new ones." What I am looking for is of
the form, "We are experts in the current standards and we see the need for
new ones to meet our requirements."
Hugh. < speaking strictly for myself >
David V James wrote:
Hugh,
Thanks for insightful throughs. We certainly agree on
plug-and-go/plug-and-play(:>), as opposed to plug-and-pray(:<).
Personally, I believe that the existing standards for provisioning,
admission control, policing and QOS could be implemented easily and
cheaply in a residential environment.
Which ones, in particular, and how would they be used?
Its hard to test such an hypothesis, without reference material
and assumptions.
If this is not the case, then proof should be presented to make
a case for new standards.
Assuming 75% time-sensitive traffic with arbitrary topology
and loading (subject the the aforementioned 75% rule), then
I believe proofs are contained within "Annex G" of:
http://www.ieee802.org3/re_study/material/index.html
Either way, I think there is a need for some architectural
specification to ensure that RE products have appropriate
capabilities to allow plug in and go operation.
Agreed! Although some of us use the term plug-and-play.
I assume (correct me if not) that these mean the same thing.