Transcript Network Effects model

Project
IEEE 802.20 Working Group on Mobile Broadband Wireless Access
<http://grouper.ieee.org/groups/802/20/>
Title
Comments on Traffic Models open issue
Date
Submitted
2005-1-17
Source(s)
Anna Tee
1301 E Lookout Dr.
Richardson, TX 75082
Voice: +1 972 761 7437
Fax: +1 972 761 7909
Email: [email protected]
Farooq Khan
1301 E Lookout Dr.
Richardson, TX 75082
Voice: +1 972 761 7929
Fax: +1 972 761 7909
Email: [email protected]
Joseph Cleveland
1301 E Lookout Dr.
Richardson, TX 75082
Voice: (972) 761-7981
Fax: (972) 761-7909
Email: [email protected]
Re:
802.20 Call for Contributions: Session # 12 – January 17-21, 2005
Abstract
Comments and suggestions are provided on one of the open issues in 802.20 traffic models.
Purpose
For discussion and adoption into 802.20 evaluation criteria document
Notice
Release
Patent
Policy
This document has been prepared to assist the IEEE 802.20 Working Group. It is offered as a basis for
discussion and is not binding on the contributing individual(s) or organization(s). The material in this
document is subject to change in form and content after further study. The contributor(s) reserve(s) the right
to add, amend or withdraw material contained herein.
The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this
contribution, and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in
the IEEE’s name any IEEE Standards publication even though it may include portions of this contribution;
and at the IEEE’s sole discretion to permit others to reproduce in whole or in part the resulting IEEE
Standards publication. The contributor also acknowledges and accepts that this contribution may be made
public by IEEE 802.20.
The contributor is familiar with IEEE patent policy, as outlined in Section 6.3 of the IEEE-SA Standards
Board Operations Manual <http://standards.ieee.org/guides/opman/sect6.html#6.3> and in Understanding
Patent Issues During IEEE Standards Development <http://standards.ieee.org/board/pat/guide.html>.
1
Evaluation of VoIP application
Summary of previous 802.20 contributions on VoIP models
 Review candidate vocoder characteristics

Variable rate: EVRC
 Fixed rate: ITU-T G.729 B

Markov models
 IP stack model
 Network effects model
 QoS metrics and performance
 Conclusion
 List of References

2
Summary of Earlier Discussions on VoIP Traffic Model

Most recent contribution: 04/37

Described Framework of VoIP model





Reviewed Vocoder standards



Voice source model
IP stack model
Network effects model
End-to-end control model
EVRC (TIA / EIA / IS-127)
ITU G.729/A (DTX - Voice Activity Detection)
Earlier contribution: 04/12



Discussed protocol overhead – TCP, UDP, RTP, PHY and MAC headers
Discussed voice activity detection
Techniques for improving capacity



Header compression
Frame aggregation
Performance metrics and requirements



Packet lost ratio
Latency
Delay jitter
3
Voice Source Model

Choose Vocoder standard used in current and foreseeable future

Candidate vocoder standards





EVRC
ITU-T G.729 A, B
ITU-T G.723.1 A
Others
EVRC Vocoder standard



TIA / EIA / IS-127
Used in 3G technology (CDMA 2000)
Variable data rates:





Full Rate: 8.6 kbps
Half Rate: 4.3 kbps
1/8th Rate: 1.0 kbps (Background noise)
Depend on average Vocoder input signal energy level
Multiple thresholds used to determine data rate based on detected energy level
4
Markov Model for Variable Rate Source Coder [IS-871]
R-1
8.6 kbps
R-½
4.3 kbps
R-1/8
1.0 kbps
R-¼
2.1 kbps
5
Markov Model for Variable Rate Source Coder [IS-871]
• State transition probabilities defined in IS-871:
• “Markov Service Options for CDMA 2000 spread spectrum system”
• Model has been used by 1xEV-DV evaluation methodology:
• Voice activity factor: 0.403
• Full rate (R-1): 29%
• Half rate (R-1/2): 4%
• Quarter rate (R-1/4): 7%
• One-eighth rate (R-1/8): 60%
6
Fixed Rate Speech Coding Standard ITU-T G.729
- Silence Compression Scheme [Annex B]
• ITU-T G.729 (CS-ACELP)
• 8 kbps
• 10 ms frame
• ITU-T G.729 A
• Reduced complexity 8 kbps CS-ACELP speech codec
• ITU-T G.729 Annex B (Silence compression scheme for G.729)
• Voice Activity Detection
• Discontinuous Transmission feature (DTX)
• Use a threshold to determine voice activity based on detected average energy level
• Two states:• 8 kbps
• 0 kbps (no transmission when inactive)
• Receiver detects DTX state to determine if:
• Active: Decode speech data
• Inactive: Generate comfort noise locally based on information on background noise level
Markov Model for Fixed Rate Source Coder with
Voice Activity Detection (DTX) [ITU-T G.729B]
P(A|I)
1-P(A|I)
1-P(I|A)
Inactive
0 kbps
Active
8 kbps
P(I|A)
• Specify values for transition probabilities:
• P(A|I)
• P(I|A)
• Assume similar Voice activity factor of 0.4
 P(A|I) ~ 0.4
 P(I|A) ~ 0.6
8
IP Stack model

Protocol used for VoIP



UDP
RTP
Evaluation criteria document does not have UDP, RTP model


Specify Header size [2]





RTP ~ 12 bytes
UDP ~ 8 bytes
IP ~ 20 bytes
RTP + UDP + IP ~ 40 bytes
Protocol header compression


Should working group decide if header compression needs to be used in the
evaluation?
If used, what should be the compression ratio?


Not really necessary except to account for the overhead
20:1 bytes ? [2]
PHY/MAC headers are proposal dependent

Proponents to specify the required header size
9
Network Effects model

Network delay model defined in 802.20 evaluation criteria document V.12

Network delay distribution



Shifted Gamma distribution
Parameters specified in Table 8 of Section 5.3.1
Network packet loss




Specified in Section 5.3.2
Assumption: 0%
Focus of evaluation is on the PHY and MAC layer technology
In comparison, the network packet loss rate is negligible assuming no congestion
10
QoS metrics for Telephony Applications

Transmission rating factor R-Value


Described in ITU-T G.175
Calculated using the E-model



Relative measurement with respect to a reference condition
Trend in transmission planning
Value range from 0 – 100 (Very high quality)

Ro ~ Basic signal-to-noise ratio
Is: combination of all impairments ~ simultaneous with voice
Id: impairment caused by delay
Ie: Equipment impairment factor ~ caused by low bit rate codecs
A: advantage factor ~ compensation of impairments in the presence of advantages

Other quality measurement metrics can be computed from the R-values







Mean Opinion Score (MOS)
Percentage of Good or Bad (GoB%)
Percentage of Poor or Worst (PoW%)
11
QoS performance measurement
– Subjective vs Objective
• Sources:
•TIA / EIA / IS-810-A
• ITU-T G.175
R-value
lower limit
MOS
Speech Transmission
Quality Category
User Satisfaction
90
4.34
Best
Very satisfied
80
4.03
High
Satisfied
70
3.60
Medium
Some users
dissatisfied
60
3.10
Low
Many users
dissatisfied
50
2.58
Poor
Nearly all users
dissatisfied
12
QoS Performance: R-value vs One-way Network Delay
• G.729A: 8 kbps, 10ms frame, Voice Activity Detection (VAD)
• “Many Users dissatisfied” when:
• Packet loss > ~ 2% (2 speech frames/packet)
• One-way Network Delay > ~180 ms (when packet loss = 0%)
13
• Source: TIA / EIA / IS-810-A
How about G.723.1 A ?
• G.723.1 A: 6.3 kbps, 30 ms frame, Voice activity detection (VAD)
• “Many Users dissatisfied” when:
• Packet loss > ~ 1% (1 speech frame/packet)
• One-way Network Delay > 110 ms (when packet loss = 0%)
14
Source: TIA / EIA / IS-810A
Conclusion

ITU-T G.729 A,B

Tolerates relatively higher packet loss and longer network delay than G.723.1A







Supports variable rate including voice activity detection
Larger number of Markov states as a result of multiple source data rate
Model for state transit described in detail in IS-871
Used in current CDMA 2000 standard
What is the R-value performance characteristics, as compared to G.729 A,B ?
Need to choose a vocoder for technology evaluation




Need to decide on the state transition probabilities
EVRC


G.729 A,B can be chosen over G.723.1 A
G.729 A, B?
EVRC?
Another vocoder standard?
Need to determine performance metrics for evaluation




Somewhat dependent on the choice of vocoder standard for evaluation
Packet loss ratio
Latency
Delay Jitter
15
List of References
J. Tomcik, “VoIP Models-Update”, IEEE C802.20-04/37, May 2004.
F. Khan, “VoIP Models for 802.20 System Performance Evaluation”, IEEE 802.2004/12, Jan 2004.
TIA/EIA, “Transmission requirements for Narrowband Voice over IP and Voice over
PCM Digital Wireline Telephones”, TIA/EIA-810-A, Dec 10, 2000.
TIA/3GPP2, “Enhanced Variable Rate Codec, Speech Service Option 3 for
Wideband Spread Spectrum Digital Systems”, TIA-127A, May 2004.
3GPP2, “Markov Service Option (MSO) for cdma2000 Spread Spectrum
Systems”, IS-871, April 2001.
ITU-T, “Coding of speech at 8 kb/s using Conjugate-Structure Algebraic-codeExcited Linear-Prediction (CS-ACELP)”, G.729, March 1996.
1.
2.
3.
4.
5.
6.


7.
8.
Annex A: Reduced complexity 8 kb/s CS-ACELP speech codec, Nov, 1996
Annex B: A Silence compression scheme for G.729 optimized for terminals conforming to
Recommendation V.70, Nov 1996.
ITU-T, “Transmission planning for private/public network interconnection of voice
traffic”, G. 175, March 1997.
“IEEE 802.20 Evaluation criteria document Ver. 12”, November 2004.
16