Re: [lmap] draft-ietf-lmap-information-model-06: Reinsert Task Capabilities and ma conditions back into the information model

["MORTON, ALFRED C (AL)" <[e-mail]>]

Hi Esteban,

The Poisson distribution determines the inter-packet sending times 
for this metric, creating a form of unbiased sampling. This follows
the IPPM Framework RFC 2330, where you can read more.

So, these are metrics & and statistics on delay measured on a stream
of packets sent according to the Poisson distribution.

You're right that the distribution of RTT will not like have a Poisson
distribution, so in that we agree.

regards,
Al

> -----Original Message-----
> From: lmap [mailto:[e-mail]] On Behalf Of Esteban Carisimo
> Sent: Tuesday, October 13, 2015 9:16 AM
> To: [e-mail]
> Subject: Re: [lmap] draft-ietf-lmap-information-model-06: Reinsert Task
> Capabilities and ma conditions back into the information model
> 
> Hello,
> 
> I joined this WG a couple a weeks ago and I have been getting involved
> with lmap's threads. Although you discussed this topic a couple a weeks
> ago, my server has been misconfigurated so far. Now I think my mail
> server is fine and I am still interested in collaborate in this topic.
> 
> When I received this email I read that your are proposing to add a
> parameter so-called 'Act_IP_UDP_Round-trip_Delay_95th-
> percentile_Poisson'. Then I read the full draft and I found that you are
> also interested in get the mean Poisson value. This information made me
> wonder if you are supposing that RTT probability density function is
> Poisson.
> 
> Maybe I misunderstood the meaning of Poisson in that parameter, if not,
> we cannot ensure that RTT probability density function is Poisson.
> Moreover, there is enough information among literature that proves that
> packet arrivals have a self-similar behaviour instead of Poisson's.
> 
> It would be interesting to discuss about RTT probability density
> function but I do not know if it concerns to this current WG and draft.
> For this reason, I think it is a very good idea to add the value of
> percetiles but I would write them uncategorized, without the Poisson
> label.
> 
> 
> PS: My apologies if you have received this email twice.
> 
> 
> Esteban Carisimo,
> Universidad de Buenos Aires and CONICET.
> 850 Paseo Colon avenue, Buenos Aires, Argentina.
> 
> 
> El 26/9/15 a las 13:51, MORTON, ALFRED C (AL) escribió:
> > Hi Tim,
> > good question (we're having Tuesday's meeting on Saturday :-(  )
> >
> > I think we need some approximate hierarchy like this:
> >
> > Measurement Capability [
> >     Measurement Protocol [
> >         Protocol Roles [ ]
> >     ]
> >     Registry URI  [
> >         Method Roles [ ]
> >     ]
> > ]
> > so, an example would be
> >
> > Measurement Capability [
> >      TWAMP [
> >         Control-Client; Session-Sender; Server; Session-Reflector;
> >         ]
> >      Prefix:Act_IP_UDP_Round-trip_Delay_95th-percentile_Poisson [
> >         Src; Dst;
> >      ... more URIs and Roles...
> >         ]
> > ]
> > for a fully capable MA.
> >
> > I think you can easily conjure-up how to indicate a system with less
> > than full capabilities, or how a different Meas. Protocols could be
> > substituted in a different meas. capability that measures the *same*
> > registered metric.
> >
> > Al
> >
> >
> >> -----Original Message-----
> >> From: Carey, Timothy (Timothy)
> >> [mailto:[e-mail]]
> >> Sent: Saturday, September 26, 2015 12:16 PM
> >> To: MORTON, ALFRED C (AL); Juergen Schoenwaelder
> >> Cc: [e-mail]
> >> Subject: RE: [lmap] draft-ietf-lmap-information-model-06: Reinsert
> >> Task Capabilities and ma conditions back into the information model
> >>
> >> Al,
> >>
> >> Indeed we do need that to be a list as you suggest but what is
> >> relationship with the role to the list of registry entries?
> >>
> >> BR,
> >> Tim
> >>
> >> -----Original Message-----
> >> From: MORTON, ALFRED C (AL) [mailto:[e-mail]]
> >> Sent: Saturday, September 26, 2015 9:12 AM
> >> To: Carey, Timothy (Timothy); Juergen Schoenwaelder
> >> Cc: [e-mail]
> >> Subject: RE: [lmap] draft-ietf-lmap-information-model-06: Reinsert
> >> Task Capabilities and ma conditions back into the information model
> >>
> >> Hi Tim and Juergen,
> >>
> >> Here's a small addition to what Tim proposed:
> >>>> *         ma-task-capability-obj  ma-supported-tasks<0..*>;
> >> It's necessary for all non-measurement tasks to have a registry
> >> providing the details of each task (and someone needs to start that
> >> development).
> >>
> >> But for measurement Tasks, there can be multiple Performance Metric
> >> Registry entries involved now, and each task will be some combination
> >> of registered metrics.
> >>
> >> So we also need ma-supported-registry-entries<0..*>;
> >> and the text below (from 3.9) needs some revisions to keep-up with
> >> the latest agreements (if I've got this right):
> >>
> >>     A Measurement Task Configuration is the same in information terms
> to
> >>     any other Task Configuration.  Both measurement and non-
> measurement
> >>     Tasks have registry entries to enable the MA to uniquely identify
> the
> >>     Task it should execute and retrieve the schema for any parameters
> >>     that may be passed to the Task.  Registry entries are specified
> as a
> >>     URI and can therefore be used to identify the Task within a
> namespace
> >>     or point to a web or local file location for the Task
> information.
> >>     As mentioned previously, these URIs may be used to identify the
> >>     Measurement Task in a public namespace
> >>     [I-D.ietf-ippm-metric-registry].
> >>
> >> Al
> >>
> >>> -----Original Message-----
> >>> From: lmap [mailto:[e-mail]] On Behalf Of Carey,
> >>> Timothy
> >>> (Timothy)
> >>> Sent: Friday, September 25, 2015 8:51 AM
> >>> To: Juergen Schoenwaelder
> >>> Cc: [e-mail]
> >>> Subject: Re: [lmap] draft-ietf-lmap-information-model-06: Reinsert
> >>> Task Capabilities and ma conditions back into the information model
> >>>
> >>> Juergen,
> >>>
> >>> But the ma-status-object doesn't provide all the conditions of the
> >>> MA might be experiencing (which might be implementation specific).
> >>>
> >>> BR,
> >>> Tim
> >>>
> >>> -----Original Message-----
> >>> From: Juergen Schoenwaelder [mailto:j.schoenwaelder@jacobs-
> >>> university.de]
> >>> Sent: Wednesday, September 23, 2015 9:52 AM
> >>> To: Carey, Timothy (Timothy)
> >>> Cc: [e-mail]
> >>> Subject: Re: [lmap] draft-ietf-lmap-information-model-06: Reinsert
> >>> Task Capabilities and ma conditions back into the information model
> >>>
> >>> On Tue, Jul 28, 2015 at 02:52:53PM +0000, Carey, Timothy (Timothy)
> >>> wrote:
> >>>> Juergen,
> >>>>
> >>>> In draft 05 we removed the ability for a measurement agent to
> >>>> report
> >>> its metrics that it supports.
> >>>> We need to add
> >>>>
> >>>> *         ma-task-capability-obj  ma-supported-tasks<0..*>;
> >>>>
> >>>> *         [ma-condition-obj        ma-conditions<0..*>;]
> >>>>
> >>>> They were in the ma-status-obj.
> >>> You are correct. There currently is no 'list' of tasks that are
> >>> supported by a device anymore and we may have to put this back in.
> >>>
> >>> We changed the reporting of runtime failures to make them more
> useful.
> >>> See the ma-task-status-obj.
> >>>
> >>> /js
> >>>
> >>> --
> >>> Juergen Schoenwaelder           Jacobs University Bremen gGmbH
> >>> Phone: [phone]         Campus Ring 1 | 28759 Bremen |
> Germany
> >>> Fax:   [phone]         <http://www.jacobs-university.de/>
> >>>
> >>> _______________________________________________
> >>> lmap mailing list
> >>> [e-mail]
> >>> https://www.ietf.org/mailman/listinfo/lmap
> > _______________________________________________
> > lmap mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/lmap
> 
> _______________________________________________
> lmap mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/lmap