Practical Considerations for Smoothing Multimedia Traffic over PacketSwitched Networks

1
Practical Considerations for Smoothing Multimedia
Traffic over Packet-Switched Networks

• Christos Tryfonas
• tryfonas_at_sprintlabs.com
• April 2002

Sprint Advanced Technology Laboratories Burlingame
, California
2
Problem Statement

• Objective Compute a rate schedule to send a
  video stream that maximizes some network
  parameter (e.g. utilization)
• Constraints Avoid decoder buffer under/overflows

3
Work-Ahead Smoothing Algorithms

• Most smoothing algorithms compute maximum
  constant rate segments
• Differences in the selection of the starting
  point of next segment

4
Outline

• Practical Observations
• Smoothing under variable network delays
• Sender-Based Smoothing
• Clock-Aware Smoothing
• Renegotiation Failure Decisions
• Conclusions

5
Practical Observations

• Network delay is variable
• Current smoothing algorithms compute schedule
  based on the receiver side
• Not all data sent by the sender is utilized by
  the receiver
• Some applications require clock to be recovered
  at the receiver (e.g. Broadcast TV)
• Consideration of Renegotiation Failures

6
Smoothing under Variable Network Delays

• Jitter usually a function of the reserved rate
  (e.g. Fair Queueing schedulers)
• Currently optimization done with a pre-specified
  worst-case network jitter, i.e., not optimal
• Overflow envelope is a function of the decision,
  not fixed
• The burstiness of the source at the selected rate
  defines the ambiguity zones

7
Our Approach

• Optimization should be done by considering the
  ambiguity zones resulting from the varying
  reserved rates
• Existing optimization algorithms should consider
  the variable ambiguity zones for their decisions

8
The Case of Rate Decrease

• New rate corresponds to higher jitter
• Lower envelope is changed to accommodate the
  increase in jitter

9
The Case of Rate Increase
Jitter 2
Schedule Ambiguity Envelope
Jitter 1 gt Jitter 2
Received Data
Peak Rate segment
Jitter 1
Time

• New rate corresponds to less jitter
• A peak-rate segment is inserted before the new
  ambiguity zone with the new jitter

10
Sender-Based Smoothing

• Not all data transmitted is inserted in the
  decoder buffer

• Two approaches
• Rate Adaptation
• Forward Schedule computation

11
Clock-Aware Smoothing
12
Renegotiation Failure Decisions

• Dynamic Requantization
• Use of scalable-encoded streams
• New Approach Pre-computed alternate schedules
  with convergence points

13
Conclusions

• Current smoothing algorithms produce infeasible
  schedules in practice
• Optimization has to consider
• Variable Network Jitter a function of the
  reserved rate
• Unused data sent by the sender but not utilized
  by the receiver/decoder
• Clock-awareness
• Renegotiation Failures
• For more ? www.sprintlabs.com/People/tryfonas