30 Jan
2006
30 Jan
'06
9:14 p.m.
Aisling -
To address your points in order:
1) It would be possible to intercept the data (RTCP) QoS packets in a
media proxy. Of course, if you want to check QoS on all your calls,
then all media must pass through this media proxy. If you come up
with some interesting way to perform this in any open-source manner,
I'm sure the community at large would be very interested in hearing
about it. I believe some Session Border Controllers have this
functionality (perhaps those vendors can chime in via private mail to
you if they have such functionality.)
2) Peer to Peer voice is more or less impossible to capture in
real-world situations without a media proxy. I don't want to make
this a rehash on why Ethernet sniffing is not viable as a production
method, but a mesh of ethernet sniffers on a switched network would
be exceedingly difficult, expensive, and sub-optimal. It would
probably be cheaper to build your own SIP endpoint devices or
software that did QoS reporting in a centralized way, or just force
the use of a media proxy.
If you are trying just to intercept a single conversation, then
Ethereal will work fine if you have access to all of the traffic on
an unswitched local segment.
JT
At 9:36 AM +0000 1/30/06, Aisling wrote:
Hi,
Thanks to John, Nils and Atle for the replies. So if I am to proceed
with extracting QoS info from RTCP it is now my understanding that I
would either have to have my QoS gathering application at:
1) the media proxy (used primarily for nat situations)
2) where the voice is going peer to peer, it would need to be placed at
the end points. However what I'm wondering about here is would it be
possible to have a QoS collector application running on another machine
on the same network as the two endpoints (obviously presuming they're
all on the same network) which could intercept the sender and receiver
reports? And if so does anyone have any idea if there's another way to
distinguish RTCP packets without knowing the IP address or port?
Many thanks,
Aisling.
-----Original Message-----
From: John Todd [mailto:jtodd@loligo.com]
Sent: 29 January 2006 22:33
To: Aisling O'Driscoll; serusers(a)lists.iptel.org
Subject: Re: [Serusers] SER RTP QoS Monitoring
Hello -
Your method #1 (RTCP) is the best solution, it would seem. Method
#2 (INFO) is sort-of supported in some devices, like the Cisco 79xx
series with SIP software - they support insertion of some basic call
element data into the BYE message. Method #3 is complex, as it would
require the SNMP collector to have full knowledge of all call states
on the devices that were queried (though possibly this is not more
complex than the RTCP collection process, either.) In any of these
three cases, SER does not collect the call quality data - the
endpoints or media proxies are the only places that see this data.
If you run a mediaproxy in conjunction with SER, then it may be
possible to collect and report on that data, but there are drawbacks
to a design of that nature.
For the long term, I would suggest taking a look at this draft,
which describes the use of RTCP-XR (or "vq-rtcpxr") as set of data to
be transmitted via SIP messages between media-handling entities and
other reporting systems:
http://www.ietf.org/internet-drafts/draft-johnston-sipping-rtcp-summary-
08.txt
While we are still a while away from seeing vendor (or open source)
implementations, this seems like a reasonable proposal. It allows
for SIP-based publication or subscription of call quality events by
various network elements, and it will permit arbitrary querying of
this data at any time interval (not just at the end of the call.)
JT
Hi,
I want to be able to extract QoS parameters such as jitter, delay and
packet loss from VoIP calls made using my SER(both peer to peer for
my public clients and also for clients where the voice will be routed
with mediaproxy). I would like to develop something as opposed to
purchasing a software package.
-------------------Legal
Disclaimer---------------------------------------
The above electronic mail transmission is confidential and intended only
for the person to whom it is addressed. Its contents may be protected by
legal and/or professional privilege. Should it be received by you in
error please contact the sender at the above quoted email address. Any
unauthorised form of reproduction of this message is strictly
prohibited. The Institute does not guarantee the security of any
information electronically transmitted and is not liable if the
information contained in this communication is not a proper and complete
record of the message as transmitted by the sender nor for any delay in
its receipt.
-------------------Legal Disclaimer---------------------------------------
The above electronic mail transmission is confidential and intended
only for the person to whom it is addressed. Its contents may be
protected by legal and/or professional privilege. Should it be
received by you in error please contact the sender at the above
quoted email address. Any unauthorised form of reproduction of this
message is strictly prohibited. The Institute does not guarantee the
security of any information electronically transmitted and is not
liable if the information contained in this communication is not a
proper and complete record of the message as transmitted by the
sender nor for any delay in its receipt.
_______________________________________________
Serusers mailing list
serusers(a)lists.iptel.org
http://lists.iptel.org/mailman/listinfo/serusers From searching the Internet the last few days I
have seen that the
following options seem to be available:
1) RTCP
RTCP stream will contain this information. The problem is that
sniffers cannot clearly distinguish RTCP packets from other UDP
packets unless something like an IP address or specific UDP port is
supplied. I had considered using TCPDUMP but then where do I position
this in the network when voice is going peer to peer between client
and I'm not sure how this can work to monitor all calls when I might
not be aware of the end user IP addresses.
2) SIP INFO extension
This also seem to be an option but most off the shelf phones don't
seem to support this and this would required modification of a SIP ua.
3) SNMP
I thought maybe SNMP might be an option but the SER snmp module no
longer exists...
Does anyone have any comments on the above? Are the statements that I
made correct or can anyone think of other ways to monitor the voice
QoS? I am trying to understand how commercial applications have
accomplished this.
Many thanks in advance for advice.
Aisling.
---- Original Message ----
From: clona(a)cyberhouse.no
To: ashling.odriscoll(a)cit.ie
Subject: Re: [Serusers] Status of SER SNMP module
Date: Sun, 29 Jan 2006 20:30:02 +0100
_______________________________________________
Serusers mailing list
serusers(a)lists.iptel.org
http://lists.iptel.org/mailman/listinfo/serusers
Hi Aisling,
The SNMP module is dead as a killed turtle :(
-Atle
* Aisling O'Driscoll <ashling.odriscoll(a)cit.ie> [060129 20:26]:
> Hello,
>
> I'm just wondering what the current status of the SER SNMP module
is?
>> Is it currently supported and is there somewhere I can find
>> documentation on it?
>>
>> Many thanks,
>> Aisling.