voip performance management - telchemy voicecon 2005 - al… · • voice over ip performance...

1 VoiceCon Fall 2005

VoIP Performance Management

Alan ClarkCEO, Telchemy

VoiceCon Fall 2005


Agenda

• Voice over IP Performance Problems– Voice quality– IP problems– Non-IP problems

• VoIP Performance Management Framework• RTCP XR• Defining Requirements• Summary


Expectations of service quality

• Listening quality– Clarity, no distracting noise/ pops/ distortion

• Conversational quality– No noticeable delay or echo

• Availability– Always available, does not drop calls

• Signaling quality– Low call setup delay, features work


Voice Quality - performance measures?

• The “mantra”– Packet Loss– Jitter– Delay

• The reality…..


Jitter

0

50

100

150

200

0 0.5 1 1.5 2

Time (Seconds)

De

lay (

mS

)

Average jitter level (PPDV) = 4.5mS


Jitter

• Jitter is strongly time varying

• Often measured using simple “Packet to Packet Delay Variation (PPDV)” metric– Only valid if jitter has a constant level– Meaningless for time varying jitter (normally the

case)

• Other metrics exist - MAPDV, Y.1541….

• Ensure that metrics used will capture information on highly variable jitter levels


Packet Loss - also time varying

0

10

20

30

40

50

30 35 40 45 50 55 60 65 70

Time (seconds)

50

0m

S A

vg

e P

ack

et

Loss

Rate

Average packet loss rate = 2.1%


Leads To Time Varying Call Quality

1

2

3

4

5

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Time

MO

S

0

100

200

300

400

500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Ban

dw

idth

(kb

it/

s)

Voice

DataHigh jitter/ loss/ discard

Codecdependant


Packet Loss & Jitter

• Packet loss is strongly time varying

• Packet loss typically occurs at the same time as jitter– Both due to network congestion

• Need to measure the length and severity of bursts of lost/ discarded packets– Burst = period of high loss/discard rate

• Discard Rate is a good measure of the impact of jitter

• Jitter is a good predictor of congestion problems


Delay

2

3

4

5

0 100 200 300 400 500 600

Round trip delay (milliseconds)

MO

S S

core

55dB Echo Return Loss

35dB Echo Return Loss

Conversational problemsEcho problems


Where does echo come from?

IP

EchoCanceller

Gateway

Echo

Round trip delay - typically 50mS+

Additional delay introduced by VoIP makes existing echo problems more obvious


Delay

• Impact of delay strongly depends on how much echo is present on the call– No echo - user can tolerate 200+mS of delay– Noticeable echo - delays over 50mS are problematic

• User perceived delay includes delays in end-equipment, which can be 50mS or more

• Therefore specifying a simple delay metric does not ensure good quality


Signal Level Problems

Temporal Clipping occurs with VAD or Echo Suppressors -- gaps in speech, start/end of words missing

Amplitude Clipping occurs -- speech sounds loud and “buzzy”

-16 dBm0

-36 dBm0


Typical VoIP SLA

Jitter < 20mS

Loss < 0.1%

Latency < 100mS

Availability 99.9%


A Better VoIP SLA

99.9% of calls/intervals haveMOS-LQ > 3.9MOS-CQ > 3.8

Degraded Service QualityEvents < 0.1/ hour[DSQ = ….]

Latency < 100mS

Availability 99.9%

Based on either referenceor actual endpoint

Also reflected in MOS-CQ

Availability of media ANDSignaling path

Transient quality problems

Based on all available data


Enterprise Scenario

IP Phones

IP Phones

IP Phones

IP VPN

Gateway

PSTN


Measuring at ISP Demarcation - active test

Active Test Functions

Test call


Active test for IP Service SLA

• Uses VoIP calls, to ensure packets are treated identically to “real” VoIP calls

• Use a Reference endpoint - I.e. a fixed configuration, known, virtual IP endpoint

• Test:– Peak times - to understand quality under load

conditions– Off-peak times - to detect problems before they

impact users


Measuring at user desktop - passive test

RTCP XR

SIP QoSReport

EmbeddedMonitoringFunction


Non-intrusive VoIP Performance Monitoring

• Most effective for end-to-end measurement

• Embedded quality monitoring function in IP endpoint

• Can measure service quality, signaling reliability……

• Collect data via RTCP XR or SIP QoS reports


VoIP Performance Management Framework

Media Path Reporting(RTCP XR)

Call Server andCDR database

VoIPEndpoint

VoIPGateway

SNMP

NetworkManagementSystem

Signaling Based QoS Reporting

Embedded Monitoring

Network Probeor Analyzer

VQ

VQ

Embedded Monitoring

VQ


RTCP XR

Loss Rate Discard Rate Burst Density Gap Density

Burst Duration (mS) Gap Duration (mS)

Round Trip Delay (mS) End System Delay (mS)

Signal level RERL Noise Level Gmin

R Factor Ext R MOS-LQ MOS-CQ

Rx Config - Jitter Buffer Nominal

Jitter Buffer Max Jitter Buffer Abs Max


RTCP XR-Based Protocols

• Media Path– RFC3611 RTCP XR -- published Nov 2003

• Signaling– SIP QoS Reporting -- last call, final by late 2005– H.323 reporting (H.460.9) -- published Apr 04– Megaco reporting (H.248.30) -- published Apr 04– MGCP -- draft, final by end 2005?

• SNMP– RTCP XR MIB -- draft, final by end 2005


Enterprise Application Using New Framework

Branch Office

IP Phone

IP VPN

IP Phone

Teleworker

IP

IP Phones

VQ

VQ VQ

VQ

VQ

VQ

VQ

VQ

Gateway

Probe

SLA MetricsNMS


Defining Requirements

• Do predeployment testing!!• Request support for RFC3611 VoIP Metrics in

IP Phones and Gateways• Use same monitoring technology in IP

endpoints, probes and analyzers• Use VoIP Performance Management

Framework• Request SLA metrics from service providers

that are meaningful for VoIP• Remember that IP problems are transient!!


Agenda

• Voice over IP Performance Problems– Voice quality– IP problems– Non-IP problems

• VoIP Performance Management Framework• RTCP XR• Defining Requirements• Summary

voip performance management - telchemy voicecon 2005 - al… · • voice over ip performance...

Documents