dhananjay kumar anna university, chennai, india … · dhananjay kumar anna university, chennai,...
TRANSCRIPT
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009
ITU-T Kaleidoscope 2009Innovations for Digital Inclusion
Dhananjay KumarAnna University, Chennai, IndiaEmail: [email protected]
DYNAMIC RESOURCE MANAGEMENT FOR DOWNLINK MULTIMEDIA TRAFFIC IN OFDMA CELLULAR
NETWORKS
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 2
Presentation Outlines
Multimedia ServicesMulti Carrier SystemsOFDMA System ModelProblem FormulationTwo Stage Rate Adaptive (TSRA) Algorithm
Bandwidth Adaptive (BWA) algorithm
Results & Discussions
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 3
Mobile Multimedia Services
Multimedia over Mobile NetworkTextImagesVideoAudio
QoS SupportDynamic & Efficient Resource Requirements
MPEG-4 video trace (Die Hard III)
Frame Size
Frame IndexCourtesy: Technical University, Berlin : http://www.tkn.tu-berlin.de/research/trace/ltvt.html
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 5
Quality Trace Compression Ratio
YUV:MP4
Frame Size Bit Rate
Mean Cov Peak/ Mean
X [kbyte] σ2/X Xmax/X^
Mean Peak
X/t[Mbps] Xmax/t[Mbps]
High Jurassic Park I
Silence of the Lambs
Star Wars IV
9.92
13.22
27.62
3.8 0.59 4.37
2.9 0.80 7.73
1.4 0.66 6.81
0.77 3.3
0.58 4.4
0.28 1.9
Medium Jurassic Park I
Silence of the Lambs
Star Wars IV
28.4
43.43
97.83
1.3 0.84 6.36
0.88 1.21 13.6
0.39 1.17 12.1
0.27 1.7
0.18 2.4
0.08 0.94
Low Jurassic Park I
Silence of the Lambs
Star Wars IV
49.46
72.01
142.52
0.77 1.39 10.61
0.53 1.66 21.39
0.27 1.68 17.57
0.15 1.6
0.11 2.3
0.053 0.94
Trace File of Some Hollywood Movie
Courtesy: Technical University, Berlin : http://www.tkn.tu-berlin.de/research/trace/pub.html
MPEG-4 Advanced Audio Coding
Courtsey: Fraunhofer IISMar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Telephone codec
Speech codec
(Speech)
Speech codec
(Music)
ULD 80 kbps
AAC-ELD64 kbps
AAC-LD64 kbps AAC-ELD
32 kbps
AAC-LC56 kbps MP3
64 kbps
HE-AAC20 kbps
Audio quality
CD like Quality (BW>15 KHz)
Wide Band (BW 7 KHz)
Narrow Band (BW 3.5 KHz)
Algorithmic delay in ms
10 20 30 40 50 100 130
AAC: Advanced Audio Coding ULD: Ultra-Low-Delay AAC-ELD: Enhanced Low Delay AACAAC-LD: Low Delay AAC
AAC-LC: Low Complexity AACHE-AAC: High-Efficiency AAC
Service classification and service parameters
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 7
User Experience Class
Service Class Service Parameters (Numerical Values)
Conversational Basic conversational service
Throughput: 20 kbit/sDelay: 50 ms
Rich conversational service
Throughput: 5 Mbit/sDelay: 20 ms
Conversational low delay
Throughput: 150 kbit/sDelay: 10 ms
Streaming Streaming Live Throughput: 2 – 50 Mbit/s
Delay: 100 msStreaming Non-Live Throughput: 2 – 50 Mbit/s
Delay: 1 sInteractive Interactive high delay Throughput: 500 kbit/s
Delay: 200 msInteractive low delay Throughput 500 kbit/s
Delay 20 msBackground Background Throughput: 5 – 50 Mbit/s
Delay: < 2s
Guidelines for evaluation of radio interface technologies for IMT-Advanced
Source: Document 8F/TEMP/568
Multi-Carrier Communication in Next Generation Networks
Higher CapacityCan approach theoretical limit
Easy support for adaptive resource Allocation
The Concept of Sub Channel/Carrier
Ability to combat inherent problems in wireless
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Diffraction
Line of sight
Scattering
Reflection
Reflection
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Multi Path Propagation
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Channel Response in Freq. Domain
f
|H(f)|
Frequency
Power
Frequency
PowerBSBS
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
The Concept of Multi-carrier
Multi-Carrier Communication Systems
MC-CDMASuitable for indoor wireless environmentFading resistance using frequency diversity Need for rake receivers
OFDMAHandling of multipath fading (ISI)Spectral efficiencyNo intra-cell interference (No Cell breathing)
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
OFDMA System Model
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Allocation information
Wireless channelto user k
Sub carrier and bit allocation scheme (M-QAM)
IFFT and parallel to serial
Add cyclic prefix
Symbos to bit mapping
Su carrier selector
FFT and serial to parallel
Remove cyclic prefix
Subcarrier 1
Subcarrier N
Subcarrier 2
User
N
1
2
Problem Formulation
),( min kkk
in NN αΦ=
kk
k ∀=∑ ,1α
⎟⎟⎠
⎞⎜⎜⎝
⎛−
−=1
5.1exp51
MSINRBER
ii
Initial subcarrier allocation:
αk is chosen such that 0 ≤ αk < 1 while satisfying the condition
To calculate Nmin , we first compute BER of the sub channel. The BER of ith subcarrier using M-QAM can be approximated by.
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Problem Formulation cont.
MNBR i
sb 2log=
⎥⎥⎥⎥
⎥
⎤
⎢⎢⎢⎢
⎢
⎡
=
∑=
kN
i
isb
kk
R
RN
1
min
The order of modulation M for the ith sub carrier ischosen such that the BER < 10-3.In M-QAM the maximum capacity could be approximated by
If Rk is the minimum bit rate needed to support an application k, and Ri
sb is the bit rate supported on the ith sub carrier, then the minimum no of required sub carrier is given by
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion
Problem Formulation cont.
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 16
ikl
ikl
L
l
K
k
N
i
ikl CINRp
l k
,,1 1 1
, )(∑∑∑= = =
α
∑∑∑= = =
=L
l
K
k
N
i sbk
ikl
av
l k
BWNR
C1 1 1
,
)(
Considering L types of applications, the sub carrier selection can be definedas optimization of
such that average capacity (bits/second per Hz), Cavrepresented below is maximum.
Development of TSRA Protocol
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 17
Rth2
Rth1
Packet/Time unit
Time Units
The BWA Algorithm
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 18
A
If Ri changes from Rth1 to
Rth2
Assign Mean ValueRti= (Rt1+ Rt2)/2
Allocate from nrt-VBR & ABR till
Rti = Ri+Riac
No No
Yes
Starts de-allocation & share among ABR &
nrt-VBR
No
Yes
If Rifluctuates Rth1≤Ri≤Rth2
Initialize Rth1, Rth2, Ri
START
Vary the modulation order to meet the
requirements
Is ΔRi is small
Monitor the change in input data rate (ΔRi) in every tmin
duration
END
Is call disconnected?
Yes
No
Is Ricentered
around Rth1or Rth2
Yes
No
Wait for tmin. Same
status ?
Yes
Wait for tmin. Ri at
Rth2?
Wait for tmin. Is ΔRi>rth?
Wait for tmin. Ri at
Rth1?
Yes
If Ri changes from Rth2 to
Rth1
A A
A
Yes Yes Yes Yes
No No
A
No
No
Average Capacity
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 19
2 4 6 8 10 12 14 161
2
3
4
5
6
7
Number of Users
Ave
rage
Cap
acity
(bit/
s/H
z)AnalyticalTSRA with BWATSRA without BWA
Throughput per OFDM Symbol
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 20
2 4 6 8 10 12 14 16
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
Number of Users
Thro
ughp
ut (M
bps)
per
OFD
M S
ymbo
l
TSRA with BWA and AMTSRA without BWA and AMTSRA with BWA without AM
Average delay for rt-VBR traffic
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 21
2 4 6 8 10 12 14 160
0.01
0.02
0.03
0.04
0.05
0.06
Number of Users
Ave
rage
Del
ay (S
econ
ds)
Average BER
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 22
2 4 6 8 10 12 14 160.5
1
1.5
2
2.5
3
3.5
4
4.5x 10
-3
Number of Users
Ave
rage
BE
R
CONCLUSION
Multi-carrier can support future multimedia servicesThe OFDMA based system offers high capacity / efficiency The TSRA algorithm can meet the QoS requirement of high quality multimedia applications while approaching theoretical limit (more than 6b/s/Hz for 128-QAM)
Mar del Plata, Argentina, 31 Aug – 1 Sep 2009 ITU-T Kaleidoscope 2009 – Innovations for Digital Inclusion 23