Channel Allocation for Smooth Video Delivery over Cognitive Radio Networks - PowerPoint PPT Presentation

Loading...

PPT – Channel Allocation for Smooth Video Delivery over Cognitive Radio Networks PowerPoint presentation | free to download - id: 7b1e8d-Y2I5O



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Channel Allocation for Smooth Video Delivery over Cognitive Radio Networks

Description:

CHANNEL ALLOCATION FOR SMOOTH VIDEO DELIVERY OVER COGNITIVE RADIO NETWORKS Sanying Li, Tom H. Luan, Xuemin (Sherman) Shen Department of Electrical and Computer ... – PowerPoint PPT presentation

Number of Views:18
Avg rating:3.0/5.0
Slides: 23
Provided by: Clar2241
Learn more at: http://bbcr.uwaterloo.ca
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Channel Allocation for Smooth Video Delivery over Cognitive Radio Networks


1
Channel Allocation for Smooth Video Delivery over
Cognitive Radio Networks
Sanying Li, Tom H. Luan, Xuemin (Sherman) Shen
Department of Electrical and Computer
Engineering University of Waterloo, Waterloo, ON,
Canada, N2L 3G1
  • Globecom 2010, FL, USA

2
Outline
  • Introduction
  • Proposed channel allocation algorithm for smooth
    video delivery over CR networks
  • System model
  • Optimization framework
  • Heuristic algorithm
  • Simulation results
  • Conclusion and future works

3
Video Streaming System
  • Video streaming demands stringent QoS support
    from the network work
  • Low delay
  • High throughput
  • Low network dynamics
  • Cognitive radio network
  • Highly dynamic due to random behaviour of primal
    users
  • Spotted wireless communication with time varying
    channel status
  • How to address the extraordinary network dynamics
    on provisioning guaranteed video quality to users?

4
Video Quality for VoD
  • Users need guarantee on the smoothness of
    playback.
  • Playout buffer is deployed to address the issue
    of network dynamics.

Playback frozen
5
Research Motivation and Goal
5
  • Motivation

How to provide guaranteed QoS for VoD within the
dynamic and resource-limited CR networks?
  • Dynamic nature of CR
  • Unpredictable PU behavior
  • Contention among SU for channel
  • Time varying channel throughput
  • QoS
  • Smooth playback without interruptions
  • Sensitive to dynamics

V.S.
  • Approach
  • Exploit the playout buffer to accommodate network
    dynamic and conduct channel resource allocation.
  • Users of different playout buffer storage can
    tolerate different levels of dynamics

6
Outline
  • Introduction
  • Proposed channel allocation algorithm for smooth
    video delivery over CR networks
  • System model
  • Optimization framework
  • Heuristic algorithm
  • Simulation results
  • Conclusion and future works

7
Network Architecture
7
(Video Source)
  • An infrastructure-based, single-hop CR network
  • N OFDM channels, N primary users
  • CR base station s allocates idle channels to M
    secondary users (CR users).
  • In total VoD users and
    best-effort (BE) users, and
  • The CR base station s delivers video flows to VoD
    users using OFDM downlink.

8
Protocol Description
8
(Video Source)
  • System works on a time-slot basis.
  • is the minimum spectrum sensing and
    transmission unit.
  • Channel allocation operates interactively at the
    interval of channel allocation epoch
  • Each epoch is composed of two phases beacon
    period and transmission period.
  • Time slot allocation based on the estimation of
    the availability of future spectrum resources and
    the current buffer storage of users.

9
QoS Requirements of VoD Users
9
  • QoS requirements of VoD users
  • Represented by the smoothness of video playback
  • The probability of playback frozen depends on4
  • Mean and variance of the inter-arrival time of
    packets
  • Mean and variance of the inter-departure time of
    packets
  • Buffer storage

4 Tom H. Luan, L. X. Cai, and X. Shen, "Impact
of network dynamics on users' video quality
analytical framework and QoS provision", IEEE
Transaction on Multimedia, 2010
10
QoS Requirements of Users
10
  • The probability of playback frozen of VoD
    user m in T seconds
  • The QoS of VoD user m is guaranteed by
  • The QoS of BE user m is guaranteed by

11
Optimization Framework
11
  • Formulate the channel allocation as an
    optimization problem

(Maximize the overall throughput)
(QoS of VoD users)
(QoS of BE users)
(One user allocated to one channel)
if channel n is allocated to user m,
Decision variable
otherwise.
12
Solution
12
  • On-OFF model of the CR channel
  • The occupancy of channel n (n 1, 2, ... N )
    follows a discrete-time Markov process.
  • p-persistent MAC of secondary users
  • Constant transmission rate of user m on channel n
    over channel allocation epoch T

Where represents the SNR of user m on
channel n
13
Heuristic Algorithm
13
  • The allocation guarantees the performance of VoD
    users.

0.01
0.05
0.02
0.1
0.05
Channel 1
0.01
0.02
0.05
Channel 2
VoD User 1
BE User 1
VoD User 2
BE User 2
VoD User 3
BE User 3
14
Simulation Results
14

M (Number of SUs) 50
(Number of VoD users) 20
(Number of BE users) 30
N (Number of channels) 5
(Channel availability) 0.3, 0.4, 0.5, 0.6, 0.7
(Transmission rate) 1000, 2000 pkts/sec
(Length of one slot) 10 ms
L (Number of slots in each channel allocation epoch) 50
(Upper bound of frozen probability) 0.05
(Lower bound of BE users mean transmission rate) 30 pkts/sec
r (Mean of video playback rate) 29.761905 pkts/sec
(Variance of video playback rate) 102
Initial buffer storage 50 pkts
Life time of VoD user 10, 30 secs
15
Compare with two schemes
15
  • Random allocation
  • SUs are randomly allocated to each channel every
    allocation epoch.
  • Greedy allocation
  • SUs are allocated to the channel with the largest
    throughput

Transmission rate of user m on channel n
Availability of channel n (not used by primary
user)
16
Change Channel Capacity
16
  • Frozen probability of VoD users

Increasing the channel capacity would decrease
the frozen probability of VoD users.
17
Change Channel Capacity
17
  • Overall throughput
  • Percentage of satisfied BE users
  • The performance of BE users is represented by the
    percentage of BE users whose throughput are
    larger than 30 pkts/sec.
  • Increasing the channel capacity would increase
    the percentage of satisfied BE users.
  • Increasing the channel capacity would increase
    the overall throughput.

18
Change the Population of VoD Users
18
  • Frozen probability of VoD users

Increasing the number of VoD users does not
change the frozen probability of VoD users.
19
Change the Population of VoD Users
19
  • Overall throughput
  • Percentage of satisfied BE users
  • Increasing the number of VoD users would decrease
    the overall throughput.
  • The performance of BE users is represented by the
    percentage of BE users whose throughput are
    larger than 30 pkts/sec.
  • Increasing the number of VoD users would decrease
    the percentage of satisfied BE users.

20
Change the Length of Allocation Epoch
20
  • Frozen probability of VoD users

Increasing L would increase the frozen
probability of VoD users.
21
Conclusion and Future Works
  • Optimal channel allocation framework
  • Exploit the user diversity in terms of the
    tolerance to the network dynamics
  • Propose a heuristic algorithm which has low time
    complexity
  • Simulation results show the heuristic algorithm
    outperforms the conventional schemes.
  • Future work
  • Investigate the optimal solution with guaranteed
    performance
  • Distributed channel allocation algorithm.

22
(No Transcript)
About PowerShow.com