Understanding the Performance ofThin-Client Gaming12011/5/11 CQR 2011 / Yu-Chun ChangYu-Chun Chang1, Po-Han Tseng2, Kuan-T...
Outline• Introduction• Experiment methodology– Experiment setup– Performance metric extraction• Performance evaluation• Co...
Introduction (1/2)3Client ServerUser’s inputsDisplay updates• Thin-client system2011/5/11 CQR 2011 / Yu-Chun Chang
Introduction (2/2)• Motivation– To understand which performance metric is moresufficient for thin-client gaming• Frame rat...
Our focus52011/5/11QoEPerf.MetricNetworkConditionServerClientThin-client programUserNetworkConditionServerCQR 2011 / Yu-Ch...
Outlines• Introduction• Experiment methodology– Experiment setup– Performance metric extraction• Performance evaluation• C...
Experiment Methodology72011/5/11
Why Do We Use Ms. Pac-Man?• Move Pac-Man to eat pills and get the score• Control through thin-client applications and move...
Ms. Pac-Man & Bot9• Ms. Pac-Man– Save score after the pacman ran out of 3 lives• Bot: ICE Pambush3 (published in IEEE CIG ...
• Three thin-client systems– LogMeIn– UltraVNC– TeamViewer• Network conditions10Network condition SettingsNetwork delay 0 ...
• Performance metric– Display frame rate– Frame distortion (MSE: Mean Square Error)• Record game play as video files in 20...
Outlines• Introduction• Experiment methodology– Experiment setup– Performance metric extraction• Performance evaluation• C...
132011/5/11Thin Clients are Different!CQR 2011 / Yu-Chun Chang
Visual Difference Really Matters!142011/5/11 CQR 2011 / Yu-Chun Chang
Statistical Regression15RegressionModelQoE(score)Independent factorsDisplay frame rateFrame distortion2011/5/11 CQR 2011 /...
Frame-Based QoE Model• Linear model• QoE =16Adjusted R-squared: 0.722011/5/11 CQR 2011 / Yu-Chun Chang
Frame-Based QoE Model17
Which Performance Metric is More Sufficient?• QoE degradation– Optimal user’s QoE – user’s QoE predicted by model• Frame r...
Frame Rate and Network Conditions19NetworkCondition2011/5/11QoEPerf.MetricServerClientThin-client programUserCQR 2011 / Yu...
The Frame Rate Prediction Model• Frame rate =• app1, app2: dummy variables– LogMeIn : app1 = 1, app2 = 0– TeamViewer : app...
The Frame Rate Prediction Model2011/5/11 21Adjusted R-squared: 0.85CQR 2011 / Yu-Chun ChangDelay of LogMeInDelay of UltraV...
Predicted Frame Rate2011/5/11 22CQR 2011 / Yu-Chun ChangNetwork delay Bandwidth
Which Thin-Client is Better?23NetworkConditions2011/5/11QoEPerf.MetricServerClientThin-client programUserCQR 2011 / Yu-Chu...
Network-Based QoE Model• QoE =2011/5/11Adjusted R-squared: 0.81
The Thin-Client with Best Performance• o symbol: empirical network condition– 300 records collected by PingER project2011/...
Conclusions & Future Work• Display frame rate and frame distortion are bothcritical to gaming performance on thin-clients•...
Thank you for your attention!2011/5/11 CQR 2011 / Yu-Chun Chang 27
Upcoming SlideShare
Loading in...5
×

Understanding The Performance of Thin-Client Gaming

2,045

Published on

The thin-client model is considered a perfect fit for online gaming. As modern games normally require tremendous computing and rendering power at the game client, deploying games with such models can transfer the burden of hardware upgrades from players to game operators. As a result, there are a variety of solutions proposed for thin-client gaming today. However, little is known about the performance of such thinclient systems in different scenarios, and there is no systematic means yet to conduct such analysis.

In this paper, we propose a methodology for quantifying the performance of thin-clients on gaming, even for thin-clients which are close-sourced. Taking a classic game, Ms. Pac-Man, and three popular thin-clients, LogMeIn, TeamViewer, and UltraVNC, as examples, we perform a demonstration study and determine that 1) display frame rate and frame distortion are both critical to gaming; and 2) different thin-client implementations may have very different levels of robustness against network impairments. Generally, LogMeIn performs best when network conditions are reasonably good, while TeamViewer and UltraVNC are the better choices under certain network conditions.

Published in: Technology, Business
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
2,045
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
0
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Understanding The Performance of Thin-Client Gaming

  1. 1. Understanding the Performance ofThin-Client Gaming12011/5/11 CQR 2011 / Yu-Chun ChangYu-Chun Chang1, Po-Han Tseng2, Kuan-Ta Chen2, and Chin-Laung Lei11Department of Electrical Engineering, National Taiwan University2Institute of Information Science, Academia Sinica
  2. 2. Outline• Introduction• Experiment methodology– Experiment setup– Performance metric extraction• Performance evaluation• Conclusion & future work2011/5/11 2CQR 2011 / Yu-Chun Chang
  3. 3. Introduction (1/2)3Client ServerUser’s inputsDisplay updates• Thin-client system2011/5/11 CQR 2011 / Yu-Chun Chang
  4. 4. Introduction (2/2)• Motivation– To understand which performance metric is moresufficient for thin-client gaming• Frame rate, frame delay, frame loss, and etc• Challenges– Most thin-client products are proprietary• Image compression, data-transmission protocol and display updatemechanism42011/5/11 CQR 2011 / Yu-Chun Chang
  5. 5. Our focus52011/5/11QoEPerf.MetricNetworkConditionServerClientThin-client programUserNetworkConditionServerCQR 2011 / Yu-Chun ChangQoEPerf.Metric
  6. 6. Outlines• Introduction• Experiment methodology– Experiment setup– Performance metric extraction• Performance evaluation• Conclusion & future work2011/5/11 6CQR 2011 / Yu-Chun Chang
  7. 7. Experiment Methodology72011/5/11
  8. 8. Why Do We Use Ms. Pac-Man?• Move Pac-Man to eat pills and get the score• Control through thin-client applications and movePac-Man in the game of server– Good network condition: score↑– Bad network condition: score↓• Score  Quality of Experience2011/5/11 CQR 2011 / Yu-Chun Chang 8
  9. 9. Ms. Pac-Man & Bot9• Ms. Pac-Man– Save score after the pacman ran out of 3 lives• Bot: ICE Pambush3 (published in IEEE CIG 2009)– Java-based controller to move the pacman– Capture the screen of the game and determine the position of thepacman, ghosts, and pillsNumber ScorePill 220 10Power pill 4 50Ghost 4 200 (after eating power pills)2011/5/11 CQR 2011 / Yu-Chun Chang
  10. 10. • Three thin-client systems– LogMeIn– UltraVNC– TeamViewer• Network conditions10Network condition SettingsNetwork delay 0 ms, 100 ms, 200 msNetwork loss rate 0%, 2.5%, 5%Bandwidth Unlimited, 600 kbps, 300 kbps2011/5/11 CQR 2011 / Yu-Chun Chang
  11. 11. • Performance metric– Display frame rate– Frame distortion (MSE: Mean Square Error)• Record game play as video files in 200 FPS112011/5/11 CQR 2011 / Yu-Chun Chang
  12. 12. Outlines• Introduction• Experiment methodology– Experiment setup– Performance metric extraction• Performance evaluation• Conclusion & future work2011/5/11 12CQR 2011 / Yu-Chun Chang
  13. 13. 132011/5/11Thin Clients are Different!CQR 2011 / Yu-Chun Chang
  14. 14. Visual Difference Really Matters!142011/5/11 CQR 2011 / Yu-Chun Chang
  15. 15. Statistical Regression15RegressionModelQoE(score)Independent factorsDisplay frame rateFrame distortion2011/5/11 CQR 2011 / Yu-Chun Chang
  16. 16. Frame-Based QoE Model• Linear model• QoE =16Adjusted R-squared: 0.722011/5/11 CQR 2011 / Yu-Chun Chang
  17. 17. Frame-Based QoE Model17
  18. 18. Which Performance Metric is More Sufficient?• QoE degradation– Optimal user’s QoE – user’s QoE predicted by model• Frame rate ismore sufficient!2011/5/11 18
  19. 19. Frame Rate and Network Conditions19NetworkCondition2011/5/11QoEPerf.MetricServerClientThin-client programUserCQR 2011 / Yu-Chun Chang
  20. 20. The Frame Rate Prediction Model• Frame rate =• app1, app2: dummy variables– LogMeIn : app1 = 1, app2 = 0– TeamViewer : app1 = 0, app2 = 1– UltraVNC : app1 = 0, app2 = 0• d: delay, l: loss rate, b: bandwidth• dl, dt, du : delay of LogMeIn, delay of TeamViewer, delay of UltraVNC2011/5/11 20CQR 2011 / Yu-Chun Chang
  21. 21. The Frame Rate Prediction Model2011/5/11 21Adjusted R-squared: 0.85CQR 2011 / Yu-Chun ChangDelay of LogMeInDelay of UltraVNCBandwidth of LogMeInBandwidth of UltraVNC
  22. 22. Predicted Frame Rate2011/5/11 22CQR 2011 / Yu-Chun ChangNetwork delay Bandwidth
  23. 23. Which Thin-Client is Better?23NetworkConditions2011/5/11QoEPerf.MetricServerClientThin-client programUserCQR 2011 / Yu-Chun Chang
  24. 24. Network-Based QoE Model• QoE =2011/5/11Adjusted R-squared: 0.81
  25. 25. The Thin-Client with Best Performance• o symbol: empirical network condition– 300 records collected by PingER project2011/5/11 25CQR 2011 / Yu-Chun Chang
  26. 26. Conclusions & Future Work• Display frame rate and frame distortion are bothcritical to gaming performance on thin-clients• LogMeIn performs the best among the threeimplementations we studied• Future work– Add more thin-clients to compare the performance– Design a generalizable experiment methodology for thin-client gaming with different game genres2011/5/11 26CQR 2011 / Yu-Chun Chang
  27. 27. Thank you for your attention!2011/5/11 CQR 2011 / Yu-Chun Chang 27

×