SlideShare a Scribd company logo
Network and Systems Laboratory
nslab.ee.ntu.edu.tw




   Te‐Yuan Huang, Kuan‐Ta Chen, Polly Huang 
            Network and Systems Laboratory
              National Taiwan University
             Institute of Information Science
                 Academia Sinica, Taiwan
                       INFOCOM, 2009
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Motivation
 Voice traffic is sensitive to network impairment
 Why VoIP sending rate is important?
   Most important factors on user satisfaction
    Sending Rate and its Variation
    High and Stable voice quality
      g                   q     y
 Why adapting sending rate is difficult?
   Aggressively?
   Conservatively?
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Motivation – Cont.
 Skype – one of the most popular VoIP software

 Q1: How Skype adapts its voice traffic?
 Q1: How Skype adapts its voice traffic?

 Q2: Is their mechanism good enough?
 Q2 I th i       h i       d      h?

 Q3: How can Skype’s policy be improved?
Network and Systems Laboratory
    nslab.ee.ntu.edu.tw




Related Work
 Skype’s voice traffic is governed by: [Bonfiglio et al.]
   Bit Rate
   Bit Rate
   Framing Time
   Redundant Data 
   Redundant Data
 PC‐PSTN calls
   G.729
   G 729
 PC‐PC calls 
   iSAC

   Only Redundant
   Only Redundant Data is controlled by Skype
                       is controlled by Skype
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Outline
 Motivation
 Related Work
 How does Skype adapt the redundancy ratio?
 How does Skype adapt the redundancy ratio?
 Is Skype’s mechanism good enough?
 How can we do better?
 Conclusion
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Outline
 Motivation
 Related Work
 How does Skype adapt its redundant data?
 How does Skype adapt its redundant data?
 Is Skype’s mechanism good enough?
 How can we do better?
 Conclusion
Network and Systems Laboratory
  nslab.ee.ntu.edu.tw




Experiment Setup
  p            p


                                   PC‐PSTN(G.729)
                                   PC PSTN(G 729)


                                      PC‐PC(iSAC)
Network and Systems Laboratory
    nslab.ee.ntu.edu.tw




Observation                                              10%
                                                       9%
G.729 (PC‐PSTN)                                      8%
  Constant bit rate                                7%
                                                 6%
  Constant framing 
  Constant framing
                                               5%
  time                                       4%
                                           3%
                                         2%
                                       1%
                                     0%                     0%
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Redundancy Ratio
         y
 Definition
   The percentage of packets that carry redundant voice 
   Th        t     f  k t  th t          d d t  i  
   data



                        4           3   2   1




            Redundancy Ratio = 2/4 = 0.5
             edu da cy at o / 0.5
the latter is used in Taiwan and Hong Kong, and the former is used more often in Mainland
                        Network and Systems Laboratory
                        nslab.ee.ntu.edu.tw




       Estimate Redundancy Ratio of G.729
        stimate Redundancy Ratio of G.7 9                                                                       10%
                                                                                                              9%
         G.729 (PC‐PSTN)                                                                                    8%
                 Constant bit rate                                                                        7%
                                                                                                        6%
                 Constant framing 
                 Constant framing
                                                                                                      5%
                 time                                                                               4%
                                                                                                  3%
                                                                                                2%
                                                                                              1%
                                                                                            0%                     0%
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Identify Redundancy Ratio
       y          y
 Redundancy Ratio of G.729 (PC‐PSTN Calls)
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Identify Redundancy Ratio
       y          y
 Redundancy Ratio of iSAC (PC‐PC Calls)
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Outline
 Motivation
 Related Work
 How does Skype adapt its redundancy ratio?
 How does Skype adapt its redundancy ratio?
 Is Skype’s mechanism good enough?
 How can we do better?
 Conclusion
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Skype s Redundancy Control Algorithm
Skype’s Redundancy Control Algorithm
 Adapt to network loss rate
 Adapt to network loss rate
 Adapt to other factors?
   Codec
   Network Loss Burstiness
   N t    kL     B ti
Network and Systems Laboratory
  nslab.ee.ntu.edu.tw




Effect of Codec

                       G.729

                                   iSAC
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Effect of Network Loss Burstiness
 G.729 (PC‐PSTN)
                                BR=2

                              BR 1
                              BR=1


                                       BR=1.5
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Outline
 Motivation
 Related Work
 How does Skype adapt its redundancy ratio?
 How does Skype adapt its redundancy ratio?
 Is Skype’s mechanism good enough?
 How can we do better?
 Conclusion
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Optimal Redundancy Control Policy
 p               y              y
 What’s the Optimal Policy?
  Minimum amount of redundancy data we 
  need to sustain the same audio quality 
  need to sustain the same audio quality
  under different network conditions
Network and Systems Laboratory
  nslab.ee.ntu.edu.tw




Emulation Flow


                                   Optimal
                                   Redundancy
                                   Ratio

                                   G.729
Network and Systems Laboratory
  nslab.ee.ntu.edu.tw




Skype vs. Optimal – G.729
  yp       p


                                   Skype




                                           Burst Ratio = 1
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Optimal Redundancy for the urst Ratio
Optimal Redundancy for the Burst Ratio
           BR=2                     BR=1.5


                                        BR=1

                                      Skype




                              G.729, MOS=3.5
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Modeling Optimal Redundancy Ratio
Modeling Optimal Redundancy Ratio
 Based on the targeted voice quality
 Take codec and burstiness into 
 consideration
  Optimal Policy for 
  G.729, targeted 
  MOS=3.5
Network and Systems Laboratory
   nslab.ee.ntu.edu.tw




Conclusion
 Explore how Skype adapts its voice traffic
   Redundancy Ratio
 Skype s policy does not factor in the individual 
 Skype’s policy does not factor in the individual
 codec and loss patterns in to consideration
 Propose a general model for optimal policy
   Consistent user satisfaction
   Extensible to general VoIP software
Network and Systems Laboratory
nslab.ee.ntu.edu.tw

More Related Content

PDF
Quantifying Skype User Satisfaction
PDF
An Empirical Evaluation of VoIP Playout Buffer Dimensioning in Skype, Google ...
PPTX
A Psychophysical Design towards Fair Bandwidth Allocation among VoIP Sessions
PPTX
PDF
Supporting Real-time Traffic: Preparing Your IP Network for ...
PPT
PDF
Text book 2 computer networks_a systems approach_peterson solution manual
PPT
ETE405-lec9.ppt
Quantifying Skype User Satisfaction
An Empirical Evaluation of VoIP Playout Buffer Dimensioning in Skype, Google ...
A Psychophysical Design towards Fair Bandwidth Allocation among VoIP Sessions
Supporting Real-time Traffic: Preparing Your IP Network for ...
Text book 2 computer networks_a systems approach_peterson solution manual
ETE405-lec9.ppt

Similar to Tuning Skype’s Redundancy Control Algorithm for User Satisfaction (20)

PDF
ETE405-lec9.pdf
PDF
CS5229 09/10 Lecture 10: Internet Routing
PDF
Eficiencia en LAN
PDF
Picturetel RSVP and Weighted Fair Queuing
PDF
An Educator's An Educator's Guide Guide
PPT
22 circuits
PDF
Gk2411581160
PDF
A dynamic performance-based_flow_control
PDF
Deep dive network requirementsfor enterprise video conferencing
PDF
ADAPTIVE
PDF
ADAPTIVE
PPTX
05 signal encodingtechniques
PDF
Comparisons of QoS in VoIP over WIMAX by Varying the Voice codes and Buffer size
PDF
Availability Computer Communication Network .pdf
PPTX
Improving service quality using Bayesian networks
PDF
[12] Nup 07 3
PDF
Broadband quality of service thirumurthy
PDF
Shah Md Zobair(063560056)
PPT
Availability or downtime of the servers can be found out
ETE405-lec9.pdf
CS5229 09/10 Lecture 10: Internet Routing
Eficiencia en LAN
Picturetel RSVP and Weighted Fair Queuing
An Educator's An Educator's Guide Guide
22 circuits
Gk2411581160
A dynamic performance-based_flow_control
Deep dive network requirementsfor enterprise video conferencing
ADAPTIVE
ADAPTIVE
05 signal encodingtechniques
Comparisons of QoS in VoIP over WIMAX by Varying the Voice codes and Buffer size
Availability Computer Communication Network .pdf
Improving service quality using Bayesian networks
[12] Nup 07 3
Broadband quality of service thirumurthy
Shah Md Zobair(063560056)
Availability or downtime of the servers can be found out
Ad

More from Academia Sinica (20)

PDF
Computational Social Science:The Collaborative Futures of Big Data, Computer ...
PDF
Games on Demand: Are We There Yet?
PDF
Detecting In-Situ Identity Fraud on Social Network Services: A Case Study on ...
PDF
Cloud Gaming Onward: Research Opportunities and Outlook
PPTX
Quantifying User Satisfaction in Mobile Cloud Games
PDF
量化「樂趣」-以心理生理量測探究數位娛樂商品之市場價值
PPTX
On The Battle between Online Gamers and Lags
PPTX
Understanding The Performance of Thin-Client Gaming
PPT
Quantifying QoS Requirements of Network Services: A Cheat-Proof Framework
PPT
Online Game QoE Evaluation using Paired Comparisons
PPTX
GamingAnywhere: An Open Cloud Gaming System
PPT
Are All Games Equally Cloud-Gaming-Friendly? An Electromyographic Approach
PPT
Forecasting Online Game Addictiveness
PDF
Identifying MMORPG Bots: A Traffic Analysis Approach
PDF
Toward an Understanding of the Processing Delay of Peer-to-Peer Relay Nodes
PDF
Inferring Speech Activity from Encrypted Skype Traffic
PDF
Game Bot Detection Based on Avatar Trajectory
PDF
Improving Reliability of Web 2.0-based Rating Systems Using Per-user Trustiness
PDF
A Collusion-Resistant Automation Scheme for Social Moderation Systems
PDF
Network Game Design: Hints and Implications of Player Interaction
Computational Social Science:The Collaborative Futures of Big Data, Computer ...
Games on Demand: Are We There Yet?
Detecting In-Situ Identity Fraud on Social Network Services: A Case Study on ...
Cloud Gaming Onward: Research Opportunities and Outlook
Quantifying User Satisfaction in Mobile Cloud Games
量化「樂趣」-以心理生理量測探究數位娛樂商品之市場價值
On The Battle between Online Gamers and Lags
Understanding The Performance of Thin-Client Gaming
Quantifying QoS Requirements of Network Services: A Cheat-Proof Framework
Online Game QoE Evaluation using Paired Comparisons
GamingAnywhere: An Open Cloud Gaming System
Are All Games Equally Cloud-Gaming-Friendly? An Electromyographic Approach
Forecasting Online Game Addictiveness
Identifying MMORPG Bots: A Traffic Analysis Approach
Toward an Understanding of the Processing Delay of Peer-to-Peer Relay Nodes
Inferring Speech Activity from Encrypted Skype Traffic
Game Bot Detection Based on Avatar Trajectory
Improving Reliability of Web 2.0-based Rating Systems Using Per-user Trustiness
A Collusion-Resistant Automation Scheme for Social Moderation Systems
Network Game Design: Hints and Implications of Player Interaction
Ad

Recently uploaded (20)

PPTX
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
PDF
Agricultural_Statistics_at_a_Glance_2022_0.pdf
PDF
Unlocking AI with Model Context Protocol (MCP)
PDF
gpt5_lecture_notes_comprehensive_20250812015547.pdf
PDF
Building Integrated photovoltaic BIPV_UPV.pdf
PDF
Review of recent advances in non-invasive hemoglobin estimation
PDF
Advanced methodologies resolving dimensionality complications for autism neur...
PDF
A comparative analysis of optical character recognition models for extracting...
PDF
Mobile App Security Testing_ A Comprehensive Guide.pdf
PDF
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
PDF
cuic standard and advanced reporting.pdf
PPTX
sap open course for s4hana steps from ECC to s4
PDF
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
PDF
Empathic Computing: Creating Shared Understanding
PDF
Reach Out and Touch Someone: Haptics and Empathic Computing
PPTX
20250228 LYD VKU AI Blended-Learning.pptx
PPTX
Programs and apps: productivity, graphics, security and other tools
PDF
Electronic commerce courselecture one. Pdf
PPT
“AI and Expert System Decision Support & Business Intelligence Systems”
PDF
Dropbox Q2 2025 Financial Results & Investor Presentation
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
Agricultural_Statistics_at_a_Glance_2022_0.pdf
Unlocking AI with Model Context Protocol (MCP)
gpt5_lecture_notes_comprehensive_20250812015547.pdf
Building Integrated photovoltaic BIPV_UPV.pdf
Review of recent advances in non-invasive hemoglobin estimation
Advanced methodologies resolving dimensionality complications for autism neur...
A comparative analysis of optical character recognition models for extracting...
Mobile App Security Testing_ A Comprehensive Guide.pdf
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
cuic standard and advanced reporting.pdf
sap open course for s4hana steps from ECC to s4
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
Empathic Computing: Creating Shared Understanding
Reach Out and Touch Someone: Haptics and Empathic Computing
20250228 LYD VKU AI Blended-Learning.pptx
Programs and apps: productivity, graphics, security and other tools
Electronic commerce courselecture one. Pdf
“AI and Expert System Decision Support & Business Intelligence Systems”
Dropbox Q2 2025 Financial Results & Investor Presentation

Tuning Skype’s Redundancy Control Algorithm for User Satisfaction

  • 1. Network and Systems Laboratory nslab.ee.ntu.edu.tw Te‐Yuan Huang, Kuan‐Ta Chen, Polly Huang  Network and Systems Laboratory National Taiwan University Institute of Information Science Academia Sinica, Taiwan INFOCOM, 2009
  • 2. Network and Systems Laboratory nslab.ee.ntu.edu.tw Motivation Voice traffic is sensitive to network impairment Why VoIP sending rate is important? Most important factors on user satisfaction Sending Rate and its Variation High and Stable voice quality g q y Why adapting sending rate is difficult? Aggressively? Conservatively?
  • 3. Network and Systems Laboratory nslab.ee.ntu.edu.tw Motivation – Cont. Skype – one of the most popular VoIP software Q1: How Skype adapts its voice traffic? Q1: How Skype adapts its voice traffic? Q2: Is their mechanism good enough? Q2 I th i h i d h? Q3: How can Skype’s policy be improved?
  • 4. Network and Systems Laboratory nslab.ee.ntu.edu.tw Related Work Skype’s voice traffic is governed by: [Bonfiglio et al.] Bit Rate Bit Rate Framing Time Redundant Data  Redundant Data PC‐PSTN calls G.729 G 729 PC‐PC calls  iSAC Only Redundant Only Redundant Data is controlled by Skype is controlled by Skype
  • 5. Network and Systems Laboratory nslab.ee.ntu.edu.tw Outline Motivation Related Work How does Skype adapt the redundancy ratio? How does Skype adapt the redundancy ratio? Is Skype’s mechanism good enough? How can we do better? Conclusion
  • 6. Network and Systems Laboratory nslab.ee.ntu.edu.tw Outline Motivation Related Work How does Skype adapt its redundant data? How does Skype adapt its redundant data? Is Skype’s mechanism good enough? How can we do better? Conclusion
  • 7. Network and Systems Laboratory nslab.ee.ntu.edu.tw Experiment Setup p p PC‐PSTN(G.729) PC PSTN(G 729) PC‐PC(iSAC)
  • 8. Network and Systems Laboratory nslab.ee.ntu.edu.tw Observation 10% 9% G.729 (PC‐PSTN) 8% Constant bit rate 7% 6% Constant framing  Constant framing 5% time 4% 3% 2% 1% 0% 0%
  • 9. Network and Systems Laboratory nslab.ee.ntu.edu.tw Redundancy Ratio y Definition The percentage of packets that carry redundant voice  Th   t f  k t  th t    d d t  i   data 4 3 2 1 Redundancy Ratio = 2/4 = 0.5 edu da cy at o / 0.5
  • 10. the latter is used in Taiwan and Hong Kong, and the former is used more often in Mainland Network and Systems Laboratory nslab.ee.ntu.edu.tw Estimate Redundancy Ratio of G.729 stimate Redundancy Ratio of G.7 9 10% 9% G.729 (PC‐PSTN) 8% Constant bit rate 7% 6% Constant framing  Constant framing 5% time 4% 3% 2% 1% 0% 0%
  • 11. Network and Systems Laboratory nslab.ee.ntu.edu.tw Identify Redundancy Ratio y y Redundancy Ratio of G.729 (PC‐PSTN Calls)
  • 12. Network and Systems Laboratory nslab.ee.ntu.edu.tw Identify Redundancy Ratio y y Redundancy Ratio of iSAC (PC‐PC Calls)
  • 13. Network and Systems Laboratory nslab.ee.ntu.edu.tw Outline Motivation Related Work How does Skype adapt its redundancy ratio? How does Skype adapt its redundancy ratio? Is Skype’s mechanism good enough? How can we do better? Conclusion
  • 14. Network and Systems Laboratory nslab.ee.ntu.edu.tw Skype s Redundancy Control Algorithm Skype’s Redundancy Control Algorithm Adapt to network loss rate Adapt to network loss rate Adapt to other factors? Codec Network Loss Burstiness N t kL B ti
  • 15. Network and Systems Laboratory nslab.ee.ntu.edu.tw Effect of Codec G.729 iSAC
  • 16. Network and Systems Laboratory nslab.ee.ntu.edu.tw Effect of Network Loss Burstiness G.729 (PC‐PSTN) BR=2 BR 1 BR=1 BR=1.5
  • 17. Network and Systems Laboratory nslab.ee.ntu.edu.tw Outline Motivation Related Work How does Skype adapt its redundancy ratio? How does Skype adapt its redundancy ratio? Is Skype’s mechanism good enough? How can we do better? Conclusion
  • 18. Network and Systems Laboratory nslab.ee.ntu.edu.tw Optimal Redundancy Control Policy p y y What’s the Optimal Policy? Minimum amount of redundancy data we  need to sustain the same audio quality  need to sustain the same audio quality under different network conditions
  • 19. Network and Systems Laboratory nslab.ee.ntu.edu.tw Emulation Flow Optimal Redundancy Ratio G.729
  • 20. Network and Systems Laboratory nslab.ee.ntu.edu.tw Skype vs. Optimal – G.729 yp p Skype Burst Ratio = 1
  • 21. Network and Systems Laboratory nslab.ee.ntu.edu.tw Optimal Redundancy for the urst Ratio Optimal Redundancy for the Burst Ratio BR=2 BR=1.5 BR=1 Skype G.729, MOS=3.5
  • 22. Network and Systems Laboratory nslab.ee.ntu.edu.tw Modeling Optimal Redundancy Ratio Modeling Optimal Redundancy Ratio Based on the targeted voice quality Take codec and burstiness into  consideration Optimal Policy for  G.729, targeted  MOS=3.5
  • 23. Network and Systems Laboratory nslab.ee.ntu.edu.tw Conclusion Explore how Skype adapts its voice traffic Redundancy Ratio Skype s policy does not factor in the individual  Skype’s policy does not factor in the individual codec and loss patterns in to consideration Propose a general model for optimal policy Consistent user satisfaction Extensible to general VoIP software
  • 24. Network and Systems Laboratory nslab.ee.ntu.edu.tw