The only portion missing is the jitter through the SBC itself, which really should be negligible. Assuming that is 0, your equation below would be correct. However, unless you are trying to place blame, the only jitter that really matters is the end to end since if the jitter buffer at the far end overflows, you have problem regardless of the origin of the jitter (OK, I am assuming you are not trying to do something with the RTP at any intermediate point). If you need to check for jitter through the SBC you should be able to take traces on each side and compare the inter-arrival timings. -Scott -----Original Message----- From: voiceops-bounces at voiceops.org [mailto:voiceops-bounces at voiceops.org] On Behalf Of Jim Dalton Sent: Thursday, October 28, 2010 8:34 AM To: voiceops at voiceops.org Subject: [VoiceOps] Question about Packet Jitter I have a question about calculating jitter. Consider the call diagram below. A SIP call flows from the source device to the Session Border Controller (SBC) to the destination device. All RTP packets are proxied through the SBC. +========+ +=====+ +=============+ | Source | Ingress Call Leg | SBC | Egress Call Leg | Destination | | Device |------------------->| |------------------>| Device | +========+ +=====+ +=============+ Jitter Src to SBC -------------------> Jitter Source to Destination ---------------------------------------------> If the packet jitter is known for the Ingress Call Leg from the source to SBC and for end to end packet flow from the source to the destination, is it possible to calculate jitter for the Egress Call Leg from the SBC to the destination device? I do not think the following relationship is accurate. (jitter Source to Destination) less (jitter Src to SBC) = (jitter SBC to Destination) Can anyone provide some guidance on this question? Thank you, Jim Dalton VoIP Least Cost Routing, Analysis, Billing 1.404.526.6053 www.TransNexus.com _______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops
On 10/28/2010 04:26 PM, Scott Berkman wrote:
However, unless you are trying to place blame, the only jitter that really matters is the end to end
Yep. Exactly. -- Alex Balashov - Principal Evariste Systems LLC 1170 Peachtree Street 12th Floor, Suite 1200 Atlanta, GA 30309 Tel: +1-678-954-0670 Fax: +1-404-961-1892 Web: http://www.evaristesys.com/
On Thu, 2010-10-28 at 16:26 -0400, Scott Berkman wrote:
The only portion missing is the jitter through the SBC itself, which really should be negligible. Assuming that is 0, your equation below would be correct.
I think I'd disagree with this. Jitter is an average measure of what's essentially noise (random variations in the arrival times of packets from a mean), and so I'd expect - for two jitter sources in series producing average jitter of j1 and j2 respectively - the overall jitter to be sqrt(j1^2 + j2^2). Intuitively, a packet which is late as a result of the first jitter source might then be further delayed or delivered early by the second - so just summing the average jitter values isn't appropriate. --Dave
-----Original Message----- From: voiceops-bounces at voiceops.org [mailto:voiceops-bounces at voiceops.org] On Behalf Of Jim Dalton Sent: Thursday, October 28, 2010 8:34 AM To: voiceops at voiceops.org Subject: [VoiceOps] Question about Packet Jitter
I have a question about calculating jitter. Consider the call diagram below. A SIP call flows from the source device to the Session Border Controller (SBC) to the destination device. All RTP packets are proxied through the SBC.
+========+ +=====+ +=============+ | Source | Ingress Call Leg | SBC | Egress Call Leg | Destination | | Device |------------------->| |------------------>| Device | +========+ +=====+ +=============+ Jitter Src to SBC ------------------->
Jitter Source to Destination --------------------------------------------->
If the packet jitter is known for the Ingress Call Leg from the source to SBC and for end to end packet flow from the source to the destination, is it possible to calculate jitter for the Egress Call Leg from the SBC to the destination device?
I do not think the following relationship is accurate. (jitter Source to Destination) less (jitter Src to SBC) = (jitter SBC to Destination)
Can anyone provide some guidance on this question?
Thank you,
Jim Dalton VoIP Least Cost Routing, Analysis, Billing 1.404.526.6053 www.TransNexus.com
_______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops
_______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops
-- David Knell, Director, 3C Limited T: +44 20 3298 2000 E: dave at 3c.co.uk W: http://www.3c.co.uk
Jitter is an average measure of what's essentially noise (random variations in the arrival times of packets from a mean), and so I'd expect - for two jitter sources in series producing average jitter of j1 and j2 respectively - the overall jitter to be sqrt(j1^2 + j2^2).
Intuitively, a packet which is late as a result of the first jitter source might then be further delayed or delivered early by the second - so just summing the average jitter values isn't appropriate.
--Dave
[JD wrote:] Very good point. The relationship of overall jitter to be sqrt(j1^2 + j2^2) makes sense. However, I am having trouble working through a proof of the math. Any suggestions of a reference on jitter that might address this? Many thanks for your insight, Jim D.
-----Original Message----- From: voiceops-bounces at voiceops.org [mailto:voiceops- bounces at voiceops.org] On Behalf Of David Knell Sent: Friday, October 29, 2010 1:37 AM To: voiceops at voiceops.org Subject: Re: [VoiceOps] Question about Packet Jitter
On Thu, 2010-10-28 at 16:26 -0400, Scott Berkman wrote:
The only portion missing is the jitter through the SBC itself, which really should be negligible. Assuming that is 0, your equation below would be correct.
I think I'd disagree with this. Jitter is an average measure of what's essentially noise (random variations in the arrival times of packets from a mean), and so I'd expect - for two jitter sources in series producing average jitter of j1 and j2 respectively - the overall jitter to be sqrt(j1^2 + j2^2).
Intuitively, a packet which is late as a result of the first jitter source might then be further delayed or delivered early by the second - so just summing the average jitter values isn't appropriate.
--Dave
-----Original Message----- From: voiceops-bounces at voiceops.org [mailto:voiceops- bounces at voiceops.org] On Behalf Of Jim Dalton Sent: Thursday, October 28, 2010 8:34 AM To: voiceops at voiceops.org Subject: [VoiceOps] Question about Packet Jitter
I have a question about calculating jitter. Consider the call diagram below. A SIP call flows from the source device to the Session Border Controller (SBC) to the destination device. All RTP packets are proxied through the SBC.
+========+ +=====+ +=============+ | Source | Ingress Call Leg | SBC | Egress Call Leg | Destination | | Device |------------------->| |------------------>| Device | +========+ +=====+ +=============+ Jitter Src to SBC ------------------->
Jitter Source to Destination --------------------------------------------->
If the packet jitter is known for the Ingress Call Leg from the source to SBC and for end to end packet flow from the source to the destination, is it possible to calculate jitter for the Egress Call Leg from the SBC to the destination device?
I do not think the following relationship is accurate. (jitter Source to Destination) less (jitter Src to SBC) = (jitter SBC to Destination)
Can anyone provide some guidance on this question?
Thank you,
Jim Dalton VoIP Least Cost Routing, Analysis, Billing 1.404.526.6053 www.TransNexus.com
_______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops
_______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops
-- David Knell, Director, 3C Limited T: +44 20 3298 2000 E: dave at 3c.co.uk W: http://www.3c.co.uk
_______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops
participants (4)
-
abalashov＠evaristesys.com -
dave＠3c.co.uk -
Jim.Dalton＠TransNexus.com -
scott＠sberkman.net