[SUSPECT ATTACHMENT] 8. H.248.1: Value of Statistics at a Termination Level; Re: H.248.1 Editing
Albrecht.Schwarz at alcatel.de
Albrecht.Schwarz at alcatel.de
Tue Jun 21 02:10:21 EDT 2005
For information only:
I've mentioned in our "Week 5" comments that ALCATEL will submitt a
contribution to next SG16 on "H.248.1 Annex E.11 Network Package -
Recommendations for Statistics Usage".
An initial proposal (file Alcatelx1 T05-SG16-050726-D H.248.1v3 nt
Package_Ed01.doc , Alcatelx1 T05-SG16-050726-D H.248.1v3 nt
Package_Ed01.pdf) is ready and uploaded
in /av-arch/avc-site/Incoming.
Prof. Okuba, could you please move the files in directory
http://ftp3.itu.int/av-arch/avc-site/2005-2008/0507_Gen/Working_Docs/
Thoughts, comments?
Regards
Albrecht
PS
Walter, I did cover your mentioned relation in clause "E.11.5.3.2.2.1".
But I didn't consider the correlation of statistics from different
Terminations of the same Context so far in the proposal. Perhaps I'll add
a
subsection on that item ("which is questionable in general for statistics
without any time information, like start/end time, or time stamps per
probe
etc; ... and I'm not aware of the existence of such kind of H.248
statistics so far ..")
"Binz, Walter"
<Walter.Binz at key To: Albrecht
SCHWARZ/DE/ALCATEL at ALCATEL, <megaco at ietf.org>
mile.com> cc:
Sent by: Subject: RE:
Statistics:nt/os, nt/or; Re: [Megaco] Call Flow Example in
megaco-bounces at i H.248.1Corrigendum1
(03/2004)
etf.org
23.05.2005 15:37
Albrecht et al.,
my understanding is that nt/or & nt/os for a 64kbps physical termination
(such as in the H.248.1 call flow example) should be about 8 times nt/dur
(8 PCM samples per ms), independent of the chosen RTP/AVP (I assume that
codec and silence suppression functions are part of the ephemeral
termination).
Therefore in the H.248.1 Call Flow Example below, nt/or for the ephemeral
termination should be about 10 times lower than nt/os for the physical
termination (and NOT the same), as a G.723.1 codec was used for the call.
I would appreciate if some guidance text can be added to the nt package.
Regards,
Walter
>-----Original Message-----
>From: megaco-bounces at ietf.org [mailto:megaco-bounces at ietf.org] On Behalf
Of
>Albrecht.Schwarz at alcatel.de
>Sent: Friday, May 20, 2005 9:12 AM
>To: megaco at ietf.org
>Subject: Statistics:nt/os, nt/or; Re: [Megaco] Call Flow Example in
>H.248.1Corrigendum1 (03/2004)
>
>
>Hello Alberto et al.,
>
>I think the question was not the *unit* of the time-based statistic
nt/dur
>versus the unit-less statistics nt/or & nt/os.
>Looking from a high-level perspective on the traffic volume related
>statistics nt/or & nt/os on circuit- and packet-side on the H.248 Context
>(MG) in this example, it looks like that the conservation law is invalid
>("traffic volume increase from C2P direction, whereas the volume is
>constant in the opposite direction P2C":-)
>Of course, whether the conservation law is applicable at all depends on
>codec type, silence suppression mode, RTP FEC mode, and other
>configurations.
>
>The crucial point is rather, which protocol layers have to be considered
>for these volume-based statistics? This basic question becomes more open
>due to the "generic scope" of these statistics: they are defined for ALL
>type of H.248 Terminations (physical T.: ALN, TDM; ephemeral T.: RTP,
UDP,
>IP, AAL2, AAL1, ATM, etc etc).
>
>This issue was frequently discussed in the past, also in last SG16
meeting.
>
>Just the note therefore that I'm currently working on a contribution on
>that subject for next SG16.
>The intent aren't any technical changes to the nt Package, I rather do
see
>a need to provide some guidance on statistics usage. Such recommendations
>may be covered in the procedural section (clause § E.11.5
>"Procedures"/H.248.1).
>
>regards
>Albrecht
>
>
>
>
>
> Christian Groves
> <christian.groves at er To: Puglisi
Alberto
><Alberto.Puglisi at italtel.it>
> icsson.com> cc:
>megaco at ietf..org
> Sent by: Subject: Re: [Megaco]
>Call Flow Example in H.248.1 Corrigendum1 (03/2004)
> megaco-bounces at ietf.
> org
>
>
> 12.05.2005 05:47
>
>
>
>
>
>Hello Puglisi,
>
>I'm not sure if I understand your question correctly. nt/dur shouldn't
have
>the
>same value as nt/os, nt/or because Duration is in Milliseconds and Octets
>Sent
>and Received are in numbers of Octets.
>
>Regards, Christian
>
>Puglisi Alberto wrote:
>
>> 22) The MGC now sends both MGs a Subtract to take down the call. Only
>> the subtracts to MG2 are shown here. Each termination has its own set
of
>> statistics that it gathers. An MGC may not need to request both to be
>> returned. A5555 is a physical termination, and A5556 is an RTP
>> termination.
>>
>>>From MGC to MG2:
>>
>> MEGACO/1 2 [123.123.123.4]:55555
>> Transaction = 50009 {
>> Context = 5000 {
>> Subtract = A5555 {Audit{Statistics}},
>> Subtract = A5556 {Audit{Statistics}}
>> }
>> }
>>
>>>From MG2 to MGC:
>>
>> MEGACO/1 2 [125.125.125.111]:55555
>> Reply = 50009 {
>> Context = 5000 {
>> Subtract = A5555 {
>> Statistics {
>> nt/os=45123, ; Octets Sent
>> nt/or=45123, ; Octets Received
>> nt/dur=40000 ; in milliseconds
>> }
>> },
>> Subtract = A5556 {
>> Statistics {
>> rtp/ps=1245, ; packets sent
>> nt/os=62345, ; octets sent
>> rtp/pr=780, ; packets received
>> nt/or=45123, ; octets received
>> rtp/pl=10, ; % packets lost
>> rtp/jit=27,
>> rtp/delay=48, ; average latency
>> nt/dur=38000 ; in millisec
>> }
>> }
>> }
>> }
>>
>> I understand that nt/os, nt/or, and nt/dur should have the same
values,
>> am I correct ?
>>
>>
>>
------------------------------------------------------------------------
>>
>> _______________________________________________
>> Megaco mailing list
>> Megaco at ietf.org
>> https://www1.ietf.org/mailman/listinfo/megaco
>
>_______________________________________________
>Megaco mailing list
>Megaco at ietf.org
>https://www1.ietf.org/mailman/listinfo/megaco
>
>
>
>
>
>_______________________________________________
>Megaco mailing list
>Megaco at ietf.org
>https://www1.ietf.org/mailman/listinfo/megaco
_______________________________________________
Megaco mailing list
Megaco at ietf.org
https://www1.ietf.org/mailman/listinfo/megaco
_______________________________________________
Megaco mailing list
Megaco at ietf.org
https://www1.ietf.org/mailman/listinfo/megaco
-------------- next part --------------
A non-text attachment was scrubbed...
Name: winmail.dat
Type: application/ms-tnef
Size: 13610 bytes
Desc: not available
URL: <https://lists.packetizer.com/pipermail/sg16-avd/attachments/20050621/cac60103/attachment-0006.bin>
More information about the sg16-avd
mailing list