An Analysis of Chaining Protocols for Video-on-Demand

1
An Analysis of Chaining Protocols for
Video-on-Demand

• J.-F. PârisUniversity of Houston
• Thomas Schwarz, S. J.Universidad Católica del
  Uruguay

2
Introduction

• Video-on-demand lets
• Different customers watch
• Different videos at
• Different times
• Very high bandwidth requirements

3
Solutions (I)

• Distributing server workload among several sites
• Content-delivery networks
• Local caches,
• Letting the server broadcast same video data to
  all customers watching the same video
• Not possible on today's Internet

4
Solutions (II)

• Let customers participate in the video
  distribution
• P2P solution
• Available distribution bandwidth grows linearly
  with the demand
• Cheap and easy to deploy
• Requires everyone to cooperate
• Must penalize selfish customers

5
Chaining

• One of the oldest VOD solutions
• S. Sheu, K. A. Hua, and W. Tavanapong. Chaining
  A Generalized Batching Technique for
  Video-on-Demand Systems. Proc. ICMS Conference,
  June 1997.
• Involves clients in video distribution process

6
Assumptions

• Customers have enough upstream bandwidth to
  forward the video to the next client
• Customer buffer sizes do not allow them to store
  entire videos
• Can only store last ß minutes
• A reasonable assumption in 1997

7
Basic chaining

• Customer requests form a chain
• First customer in the chain receives its data
  from the server
• Subsequent customers receive their data from
  their immediate predecessor
• Chain is broken each time two consecutive
  requests are more than ß minutes apart

8
An example
Stream from server
Customer A
Stream from customer A
Customer B
Customer C
Stream from server
9
Expanded chaining

• Assumes that
• Customers have enough buffer space to cache the
  whole contents of the video
• Helps with rewind command
• Customers will disconnect once they have finished
  playing the video
• A realistic assumption

10
How it works
From server
11
Server bandwidth requirements (2-hour video)
12
Accelerated chaining

• Has clients forward their video data to the next
  client in the chain at a slightly higher rate
  than the video consumption
• Acceleration factor will vary between 1.01 and 1.1

13
How it works
From server
Customer A
From server
Customer B
From A
From server
Dt
From B
Customer C
Dt
SERVER
To A
ToB
To C
14
Server bandwidth requirements (2-hour video)
15
Motivation for further work

• All these results were obtained through discrete
  simulation
• Mere numerical values
• Could we not use analytical methods?
• Would get algebraic solutions
• Could derive maxima/minima

16
Our assumptions

• D is video duration
• ß is buffer size
• ? is customer arrival rate
• f is video acceleration rate
• Time between arrivals is governed by the
  exponential distribution with probability density
  function
• p(t) ? e-?t

17
Basic chaining (I)

• Two cases to consider
• Interarrival time is less than ß
• Previous customer forwards the video
• No server workload
• Interarrival time is more than ß
• Server transmits whole video

18
Basic chaining (II)

• Average server workload per video is
• Average server bandwidth is

19
Expanded chaining (I)

• Two cases to consider
• Interarrival time ?t is less than D
• Previous customer forwards part of the video (D
  ?t)
• Server transmits remaining part (?t)
• Interarrival time ?t is more than D
• Server transmits whole video

20
Expanded chaining (II)
Customer A
Customer A
First case Customer B
From A
Dt
Second caseCustomer C
From server
21
Expanded chaining (III)

• Average server workload per video is
• Average server bandwidth is

22
Accelerated chaining (I)

• Two cases to consider
• Interarrival time ?t is less than D
• Previous customer forwards part of the video
• min(D, f (D ?t))
• Server transmits remaining part
• Interarrival time ?t is more than D
• Server transmits whole video

23
Accelerated chaining (II)

• Result is a fairly complicated expression
• with ? 1/f

24
Comparing analytical results with simulation
results (I)
25
Comparing analytical results with simulation
results (II)
26
Conclusion

• Very good agreement between analytical and
  simulation results
• Two techniques validate each other
• Analytical results provide a better investigation
  tool than simulation results
• Can compute bandwidth maxima,

27
Future work

• Add an incentive mechanism
• To penalize freeloaders
• Investigate how mechanism interacts with protocol
• Implement fast forward/jump
• Develop a test bed implementation

28
Handling early terminationOriginal schedule
29
Handling early terminationAfter customer B
leaves
From server
Customer A
Already played
Customer B
From server
Dt


Customer C
Dt
SERVER
To A
To C