Upload
xandra-hamilton
View
15
Download
2
Embed Size (px)
DESCRIPTION
Better-Behaved Better-Performing Multimedia Networking. Jae Chung and Mark Claypool (Avanish Tripathi) Computer Science Department Worcester Polytechnic Institute. Outline. Motivation Approach Experimental Setup Results Conclusion Future Work. Motivation. - PowerPoint PPT Presentation
Citation preview
Better-Behaved Better-Performing Multimedia Networking
Jae Chung and Mark Claypool(Avanish Tripathi)Computer Science DepartmentWorcester Polytechnic Institute
Outline
MotivationApproachExperimental SetupResultsConclusionFuture Work
Motivation
The Internet is predominantly designed for TCP traffic.
Best Effort ( no timing constraints) Reliable delivery
TCP is unsuitable for Multimedia applications…Multimedia applications use UDP
Illustrative Example
Approach
But, UDP is unresponsive to congestion...Model the TCP congestion control strategy for UDP flows Exponential decrease in data rate when
congestion is encountered Linear increase in data rate in the absence
of congestion [ Jacobson et al. 1988]
Issues
Multimedia Scaling to enable the variations in the data
rate
Congestion detection mechanism when to reduce the data rate
Maximum Frame Rate (Scale 4) = 30 frame/sec
ScaleTransmission Policy
(Pattern 1)Estimated AverageTrans. Rate (Kbps)
4 I B B P B B P B B I 1056
3 I B P B P B I 896
2 I P P I 736
1 I P I 544
0 I I 352
Directly associate transmission rates to scale valuesMPEG_APP : based on MPEG-I encoding [Walpole et al. 1997]
Multimedia Scaling
Congestion Detection
Frame loss is the indicator for congestion.A frame is assumed lost when : the receiver gets a frame whose sequence
number is greater than the expected sequence number.
the receiver does not get any frame within a timeout interval.
Operation
In case of congestion reduce scale value by half ( exponential decrease)In the absence of congestion increase the scale value by one (linear increase)
ExperimentsNS - WAN Simulator developed at University of California, Berkeley
Simulation
Time ( s)
5 TCP Flows
MM UDP Flow
MPEG UDP Flow
5 2515
MM UDP Flow
MPEG UDP Flow
TCP Flow
Simulation Results
Average Per-flow Throughput ( Light Load)
827.8750.5
839.2
1047.2
0.0
200.0
400.0
600.0
800.0
1000.0
1200.0
Flow-Controlled MM Unresponsive MM
Kb
ps FTP-TCP
MM-UDP
Simulation Results
Average Per-flow Throughput ( Heavy Load)
455.0
300.2
717.7
973.6
0.0
200.0
400.0
600.0
800.0
1000.0
1200.0
Flow-Controlled MM Unresponsive MM
Kb
ps FTP-TCP
MM-UDP
Simulation Results
Conclusions
“TCP-like” multimedia protocol built upon UDP
Responsive to congestion Better Multimedia Quality
Quantitative comparison of TCP, UDP and responsive UDPImplementation in NS
Future Work
Content based multimedia scalingExperiments with other media formatsEvaluation of perceptual quality by user studiesQuantitative analysis of data loss ( ratio of packets sent and packets lost)“TCP Friendly” Evaluation