05 wo np02 e1_1 umts capacity estimation-64
TRANSCRIPT
UMTS Capacity Estimation
ZTE University
Content
UMTS Service mode
Common Capacity Design Methods
Uplink Capacity Estimation
Downlink Capacity Estimation
Estimation Examples
CS Domain Service Model
Key parameter: call frequency, call duration,
blocking probability
Average Erlang = call frequency ×duration /
3600
Call DurationCall Duration Call DurationCall Duration
Call SetupCall Setup Call ReleaseCall Release Call SetupCall Setup Call ReleaseCall Release
PS Domain Service Model
Session (WWW)Session (WWW) Session (WWW)Session (WWW)
CallCall CallCall Call (Web Page)Call (Web Page)
ClickClick ClickClick ClickClick
ActiveActive ActiveActive ActiveActive
DormantDormant DormantDormant
PacketPacket PacketPacket PacketPacket
ActiveActive
PS Domain Service Model
Dormant status and Active status conversion
Every session can contain several packet calls,
different data services and different user types
have different features
Resource occupied by packet call varies alone
with the burst transmission
PS Service Model - Example
service Bearer
rate(k
bps)
Mean
packet
size(byte)
Mean
packets
in a call
Mean
calls/s
ession
Reading
time
between
calls(second)
Email 64 480 32 2 5
www 144 480 25 5 5
Download 64 480 62 2 5
MMS 64 480 32 2 5
Streaming 384 480 267 1 0
Parameter Name Parameter definition Unit
DL Bit rate Downlink service bit rate kbps
DL Mean Packet Size Mean downlink packet
size
Byte
DL Mean # Packets Mean downlink packet
quantityDL Mean Calls/session Mean calls of downlink
sessionDL Reading time between
callsTransmission duration
between downlink calls
second
DL Mean packets in a call Mean packets in one
downlink sessionDL BLER Downlink service quality
requirementDL PS Activity Factor Downlink activating factor
PS Domain Service Model
UL Bit rate Uplink service bit rate kbps
UL Mean Packet Size Mean uplink packet size Byte
UL Mean # Packets Mean uplink packet
quantityUL Mean Calls/session Mean calls of uplink
sessionUL Reading time between
callsTransmission duration
between uplink calls
second
UL Mean packets in a call Mean packets in one
uplink sessionUL BLER Uplink service quality
requirementUL PS Activity Factor Uplink activating factor
BHSA Busy hour sessions
attempt
PS Domain Service Model
Service Category
Service type Basic characteristic Example
Conversation The time relationship
between information entities
in the stream must be kept,
session mode (small delay,
strict delay jitter requirement)
Voice, video phone
Streaming The time relationship
between information entities
in the stream must be kept
Multimedia data
stream
Interactive Request/response mode,
data integrity must be kept
Web browser,
internet game
Background Data integrity must be kept,
high delay tolerance
Email download in
background
User Group Classification
Classification principle
Based on user consumption capability and consumption behavior
Note: User groups are distinguished by service type, service rate, service quality and service intensity.
User type Group features
High-end High income group, enterprises and managers.
Providing high rate access service.
Medium-end General enterprises and some high income consumers.
Providing information inquiry, mobile entertainment and
mobile financial services.
Lower-end Middle income class and students. Providing data
services such as SMS and some mobile game services
Service Penetration
Percentage of user distribution in different application
environments are different
Percentage of high-end, middle-end and lower-end users
in different application environments are different
Service model statistic characteristic relates to
percentages mentioned above
A B C D
Total 10% 30% 30% 30%
High End 30% 10% 5% 0%
Medium End 40% 50% 40% 10%
Low End 30% 40% 55% 90%
Traffic Analysis for Single Subscriber
CS Domain
Mean busy hour Erl. Per user=mean busy hour
calls*mean call duration/3600
Service
type
Mean busy
hour calls
Mean
call
duration
Activate
factor
Mean
speed
(kbps)
Mean busy hour
erl per user
Tel. 1.25 72 0.5 12.2 0.025
Video
phone
0 (lower end)
0.05 (medium
end)
0.1 (high end)
54 1 64 0 (lower end)
0.00075 (middle
end)
0.0015 (high end)
Traffic Analysis for Single Subscriber
PS Domain
Node: penetration rate means the percentage of UEs which support this service in total UEs.
Busy hour throughput per user = BHSA* mean calls in a session *mean packets in a call*mean packet size*8/1000
Equivalent Erl = Busy hour throughput per user / (Bearer rate *3600)
Service type Penetration
rate
BHSA Mean
packet
size
(byte)
Mean
packets
in a call
Mean
calls/s
ession
Busy hour
throughput
per user
(kbit)
Web
service
Low-end
user
50% 0.01 480 25 5 4.8
Medium
end user
75% 0.02 480 25 5 9.6
High-end
user
100% 0.03 480 25 5 14.4
Traffic Analysis for Single Subscriber
The average traffic according to the Service Model
in each transmission environment is :
Average traffic for each subscriber = ∑ Ratio of
subscriber group* Service penetration * average
traffic of this group
Content
UMTS Service mode
Common Capacity Design Methods
Uplink Capacity Estimation
Downlink Capacity Estimation
Estimation Examples
Input:system load requirment and
coverage requirement
Uplink coverage
estimation
Quantity of BSs
satisfying uplink
coverage
Downlink coverage
estimation
Quantity of BSs
satisfying downlink
coverage
Compare the results
and evaluate the
larger one
Uplink capacity
estimation
Quantity of BSs
satisfying uplink
capacity
End
Based on traffic type Based on power
Quantity A of
channels to be
provided by every cell
on the downlink
Quantity B of
channels availably
provided by every
cell on the downlink
Add B
Ss
No
Yes
A<B
Dow
nlin
k ca
paci
ty
estim
atio
n
UMTS Network Dimensioning Procedure
Capacity Estimation Procedure
Hybrid service intensity analysis
The UMTS system provides multiple services and the hybrid
service intensity analysis makes the system capacity consumed by
various services equivalent to that consumed by a single service.
Uplink capacity estimation
Estimate the NodeB number that meets the service demand based
on the hybrid service intensity analysis.
Downlink capacity estimation
It is a verification process. The NodeB transmission power formula
is used to calculate the channel number that can be provided by
the current NodeB scale so as to verify whether this channel
number can meet the capacity requirement, and if it cannot,
stations need be added.
Common Capacity Design Methods
Equivalent Erlangs method
Post Erlang-B method
Campbell method
Equivalent Erlangs Method
Principle: Make a service equivalent to another service and
calculate the total Erl.
Example
Service A: 1 channel for each connection and the total is 12 erl.
Service B: 3 channels for each connection and the total is 6 erl.
If 1 erl service B = 3 erl service A, altogether 30 erl service A shall
be equivalent and 39 channels shall be required (under 2%
blocking rate).
If 3 erl service A = 1 erl service B, altogether 10 erl service B shall
be equivalent and 17 service B channels shall be required (equal
17*3=51 service A channels under 2% blocking rate).
+
Low speed
service
equivalent
High speed service
equivalent
2 Erl low
speed
service
1 Erl high
speed service
Capacities meeting the same GoS are different
The calculation result is related to the equivalent mode
Equivalent Erlangs Method
Post Erlang-B Method
Principle: Calculate the capacity required by each service respectively and add them.
Example Service A: 1 channel for each connection and the total
is 12 erl.
Service B: 3 channels for each connection and the total is 6 erl.
Service A requires 19 channels (under 2% blocking rate).
Service B requires 12 service B channels (equal 12*3=36 service A channels, under 2% blocking rate).
These two services require 19+36=55 channels
Post Erlang-B Method
Suppose services A and B are the same kind, where, Service A: 1 channel for each connection and the total is 12 erl.
Service B: 1 channel for each connection and the total is 6 erl.
Based on the Post Erlang-B method Service A requires 19 channels (under 2% blocking rate).
Service B requires 12 channels (under 2% blocking rate).
Altogether 19+12=31 channels are required.
Based on traditional Erlang-B method
The total traffic of services A and B is 12+6=18 erl and altogether 26 channels are required under 2% blocking rate.
Required channel number estimated through the Post Erlang-B method is too large.
1 Erl service A
1 Erl service B
+
1 Erl service A and
1 Erl service B
Capacities meeting the
same GoS are different
The calculation
result is too
pessimistic
Post Erlang-B Method
Campbell Method
Principle: Make multiple services equivalent to a virtual
service and calculate the capacity on the basis of the
virtual service.
c
aCCapacity ii
cfficOfferedTra
i
ii
i
ii
aerl
aerl
c
2
ic
iserviceofcapacityCiiserviceofamplitudea
niancevnmeana
factorcapacityc
i
......
*var*
.
3036112 ii aerl
6636112 222ii aerl
Campbell Method
Example
Service A: 1 channel for each connection and the total is 12 erl.
Service B: 3 channels for each connection and the total is 6 erl
Mean & variance
2.230
66
c
63.132.2
30 Traffic Offered c
α
Campbell Method
Capacity factor c
Virtual traffic
21 channels (virtual channels) are required to
meet the virtual traffic under 2% blocking rate.
471)2.221(1 C
493)2.221(2 C
Campbell Method
Under 2% blocking rate, channel number required by each
service is shown as follows:
Service A:
Service B:
Different channel numbers are required to meet the GOS
requirements of diversified services.
Compared with the former two methods, the calculation
result through the Campbell method is more reasonable.
Campbell Method
If the reference service is the voice service:
voicevoicevoice
serviceserviceservice
servicevNoEbR
vNoEbRAmplitude
*/*
*/*
Content
UMTS Service mode
Common Capacity Design Methods
Uplink Capacity Estimation
Downlink Capacity Estimation
Estimation Examples
jtotal
j
jjj
PI
P
Rv
WNoEb
)/(
W: indicates the chip rate.
vj: indicates user j’s activation factor.
Rj: indicates user j’s data rate.
Pj : indicates user j’s signal receive power
Itotal: indicates total broadband receive power with
the thermal noise power included in the NodeB.
Uplink Load Analysis
Eb/No the receive signal in the NodeB must reach
Eb/No required by the service demodulation.
totaljtotal
jjj
j ILI
vRNo
Eb
WP
)(1
1
jjj
total
jj
vRNo
Eb
WI
PL
)(1
1
N
j
totalj
N
j
j ILP11
Uplink Load Analysis
The receive power at the NodeB receive end should meet the following formula so that the user signal can meet the demodulation requirement:
Define a connection load factor Lj:
The total receive power of all N users from one cell is:
Uplink Load Analysis
The total receive power at the NodeB receive end
consists of three parts:
Neighbor cell’s interference factor I
i= Other cell interference /Local cell interference
Notherintatal PPPI
indicates the total interference power of in-cell users.
indicates the total interference power of out-cell users.
indicates the NodeB thermal noise power.
inP
otherP
NP
N
j
tataljotherin ILiPP
1
)1(
N
j
jotherintatal
tatal
N
total
LiPPI
I
P
INR
1
)1(1
1
Uplink Load Analysis
The total user receive power of the NodeB:
Define the noise lifting as the ratio of total
broadband receive power to the noise power of
the NodeB:
N
j
jjj
N
j
jUL
vRNoEb
WiLi
11
)/(1
1)1()1(
UL
NR
1
1
)1(10)( 10 ULLOGdBNR
Uplink Load Analysis
Define the uplink load factor to be:
The noise lifting can be represented to be:
25 30 35 40 45 50 55 60 65
2
3
4
5
6
7
8
9
10
11
user number
nois
e ris
e(dB
)
Shanghai dialectMinnan
dialect
mandarin
Cantonese
Uplink Load Analysis
The uplink capacity is limited by interference
increase:
Uplink Capacity Estimation
In the case of a single service, evaluate the channel
quantity provided by every cell according to the load
formula and further evaluate the total number of base
stations satisfying the uplink capacity requirement.
To budget composite traffic, based on the Campbell
algorithm, make different services consumption on the
system resource equivalent to the single service
consumption on the system resource, and then evaluate
the quantity of channels to be provided by every cell
according to load formula, and further evaluate the number
of base stations satisfying the composite traffic
requirement.
R99/HSUPA mixed calculation
During the uplink capacity calculation ,decide how
much uplink load will be designed in R99 and
HSUPA
By simulation, calculate how much PS throughput
can be carried by HSUPA
Calculate how much of the remaining PS service
to be carried by R99
Calculate equivalent
intensity of services
Calculate the variance, average value and
capacity factor of the composite service
System virtual traffic A
Calculate the quantity of
equivalent voice channels
in the cell
Quantity of virtual
channels in the cell
Virtual service capacity
B of the cell
Number of
cells
A/B
R99 Uplink Capacity Algorithm
Content
UMTS Service mode
Common Capacity Design Methods
Uplink Capacity Estimation
Downlink Capacity Estimation
Estimation Examples
Downlink Load Analysis
To correctly demodulate useful signals, the UE must
overcome interference from the following three aspects
Nothertatal PPPI )1(
P represents total power of signals from current cell
represents total interference power of signals
from the outside of the cell
represents thermal noise power from the UE
represents orthogonal factor of the downlink
otherP
NP
Downlink Load Analysis
By referring to the derivation means of uplink load
factor, denote the downlink load factors as follows:
N
j j
j
jDL iRW
NoEbv
1
])1[(/
)/(
W represents chip rate at 3.84M chip/s
vj represents activation factor of the user j
jR represents bit rate of the user j
represents the average orthogonal factor in a cell
irepresents the average ratio of the NodeB power from
other cell to that from this cell
Downlink Load Analysis
Total downlink power allocation
DL
N
jj
j
b
jMS
TXBS
RW
NE
LWN
P
1
1
0
_
Where, represents the noise power spectrum density
on the front of the receiver in the mobile station
represents the average path loss of the cellL
)(suppose KTNFNFKTNMS 290174 =+-+
MSN
46 48 50 52 54 56 58 60 62 64
32
34
36
38
40
42
44
46
user number
Tx P
ow
er
(dB
m)
Public channel
Two users
One user
Three users
.
.
.
Downlink
power
Downlink Load Analysis
The downlink capacity is limited by transmission
power of the base station
Downlink Load and Scale Analysis
Estimate downlink capacity after analyzing the
channel quantity required by uplink capacity, and
observe whether the downlink can support the
mobile station to work in the designated coverage
area and its channel quantity reaches the channel
quantity generated by the uplink
Calculate the quantity of equivalent voice
channels to be provided by every cell
Calculate the quantity of equivalent voice
channels availably provided by every cell
Compare the above two results
Content
UMTS Service mode
Common Capacity Design Methods
Uplink Capacity Estimation
Downlink Capacity Estimation
Estimation Examples
Assumed Conditions
Channel environment: downtown area TU 3 km/h
System design load: 50%
Voice service blocking rate: 2%
Interference factor from the adjacent cell: 0.65
Area of the city zone: 40.8 square kilometers
Voice CS64 PS64/64 PS64/128 PS64/384
Data rate(k) 12.2 64 64 64 64
Activity factor 0.67 1 1 1 1
Eb/No 4.2 2.87 1.6 1.6 1.6
Forecast traffic 3000 400 100 5 2
Voice CS64 PS64/64 PS64/128 PS64/384
Data rate(k) 12.2 64 64 128 384
Activity factor 0.58 1 1 1 1
Eb/No 7.7 7.7 7.4 6.4 8
Forecast traffic 3000 400 100 35 20
Uplink:
Downlink:
Assumed Conditions
Input: system load requirement and coverage requirement
Uplink coverage
estimation
Downlink coverage
estimation
Uplink capacity
estimation
Quantity of base stations
satisfying uplink coverage
Quantity of base stations
satisfying coverage
requirement
Quantity of base stations
satisfying downlink coverage
Quantity B of channels
provided by the cell
Compare the results and evaluate the larger one
End
Quantity A of channels
required by the cell
A<BAdd base stations
Based on traffic
modelBased on power
Yes
No
Estimation Flow Chart
Emission
end
Maximal emission power
(dbm)
Antenna gain (dbi)
Human body loss (db)
Effective emission power
Receiving
end
Thermal noise power
spectrum density (dbm/HZ)
Thermal noise power (dbm)
Receiver noise
coefficient (db)
Receiver noise (dbm)
Interference margin (db)
Bit rate (kbit)
Processing gain (db)
Receiving Eb/No (db)
Receiver sensibility
Antenna gain (dbi)
Line loss
Other
Power control margin
Soft handoff gain
Shade fading margin
Penetration loss
Maximal path loss
Uplink Coverage Estimation
1. Uplink budget
2. Calculate the cell coverage radius based on a specific propagation model:
Path loss k1 k2log(d) k3Hms k4log(Hms) k5log(Heff) +
k6log(Heff)log(d) k7(diffraction loss) clutter loss
30Heff-6.55k6
-13.82k5
44.6k2
152.4k1
0.540.540.540.50.65Radius
(Km)
PS64/384PS64/128PS64CS64Voice
Uplink coverage is limited by the CS64 kps service
Uplink Coverage Estimation
3. Calculate the quantity of base stations required by uplink
Coverage area of the three-sector base station
22 488.05.05.095.138
9KmRS
The quantity of base stations is 40.8/0.488=84
Uplink Coverage Estimation
af amplitudefor 1 amplitudefor ratebit
servicefor servicefor ratebit
amplitude Relative
0
0
NE
NE
b
b
Voice: 1
CS64: 64 x 1 x 100.287/12.2 x 0.67 x = 5.76
PS64/64: 64 x 1 x 100.16/12.2 x 0.67 x = 4.3
PS64/128: 64 x 1 x 100.16/12.2 x 0.67 x = 4.3
PS64/384: 64 x 1 x 100.16/12.2 x 0.67 x = 4.3
Equivalent
intensity of
each service
Variance, mean and
capacity factor of the
composite service
Virtual
traffic A of
the system
Quantity of
equivalent
voice
channels in
the cell
Quantity of
virtual
channels in
the cell
Number of
cells
A/B Virtual
traffic A of
the cell
Equivalent intensity of each service
Uplink Capacity Estimation
42.010
42.010
42.010
42.010
i
iiaerlmean 1.57663.423.453.410067.540013000
i
iiaerliance 7.182713.423.453.410067.540013000var 2222
Mean
Virtual traffic of the system mean/capacity factor 5766.1/3.17 1818.96(Erl)
Equivalent
intensity of
each service
Variance, mean and
capacity factor of the
composite service
Virtual
traffic A of
the system
Quantity of
equivalent
voice channels
in the cell
Quantity of
virtual channels in
the cell
Number of
cells
A/B
Virtual traffic
A of the cell
Capacity factor variance/mean 3.17
Variance
Uplink Capacity Estimation
Quantity of equivalent voice channels availably
provided by the cell
N
j
o
bj
N
EvR
Wf
1*
1*1
1*)1(
%50 65.0f
Get the quantity of equivalent voice channels N 54
Where, and
Equivalent
intensity of
each service
Variance, mean and
capacity factor of the
composite service
Virtual
traffic A of
the system
Quantity of
equivalent
voice channels
in the cell
Quantity of
virtual channels in
the cell
Number of
cells
A/B
Virtual
traffic A of
the cell
Uplink Capacity Estimation
Equivalent
intensity of
each service
Variance, mean and
capacity factor of the
composite service
Virtual
traffic A of
the system
Quantity of
equivalent
voice channels
in the cell
Quantity of
virtual channels in
the cell
Number of
cells
A/B Virtual traffic
A of the cell
Quantity of virtual channels in every cell
c
aCCapacity ii )(
Quantity of virtual channels in the cell (54 1)/3.17 16
Virtual traffic of every cell
Look up the Erl B table, and provide 9.83Erl for 16 virtual
channels in the case of 2% of call loss ratio
Uplink Capacity Estimation
Equivalent
intensity of
each service
Variance, mean and
capacity factor of the
composite service
Virtual
traffic A of
the system
Quantity of
equivalent
voice channels
in the cell
Quantity of
virtual channels in
the cell
Number of
cells
A/B Virtual traffic
A of the cell
Uplink Capacity Estimation
Number of cells=Virtual traffic of the system/virtual
traffic of every =1818.96/9.83=186
Number of three-sector base stations=186/3=62
Downlink Capacity Estimation
Integrate uplink and downlink coverage budget and uplink capacity
budget to determine that there are 84 base stations currently and
authenticate whether downlink power meets the requirement.
Quantity A of channels to be
provided by the cell
Average traffic of every
cell
Virtual traffic of every cell
Quantity of virtual
channels in every cell
Determine the number of
stations
Qu
antity
B o
f chan
nels av
ailably
pro
vid
ed b
y th
e ce;;
End A<BYes NO
Ad
d b
ase station
s
Voice: 3000/84/3 11.9 Erl
CS64: 400/84/3 1.59 Erl
PS64/64: 100/84/3 0.4 Erl
PS64/128: 35/84/3 0.14 Erl
PS64/384: 20/84/3 0.079 Erl
Downlink Capacity Estimation
Average traffic of various
services in every cell
Determine the number
of stations
Quantity A of
channels to be
provided by the
cell
Average
traffic of
every cell
Virtual traffic of
every cell
Quantity of
virtual channels
in every cell
Qu
antity
B o
f chan
nels av
ailably
pro
vid
ed b
y th
e cell
End
A<B
Yes
Virtual traffic of every cell
Equivalent service intensity of each service on
the downlink
Voice: 1, CS64: 7.8, PS64/64: 7.3
PS64/128: 13.1, PS64/384: 50
Mean of composite traffic
Variance of composite traffic
355.19=50 0.079+13.1 0.14+7.3 0.4+7.8 1.59+ 11.9var 2222 iance
Traffic factor capacity factor variance/mean
355.19/33.04 10.75
Virtual service capacity of the cell mean/capacity factor
33.04/10.75 3.07 (Erl)
33.04=50 0.079+13.1 0.14+7.3 0.4+7.8 1.59+ 11.9 mean
Determine the number
of stations
Quantity A of
channels to be
provided by the
cell
Average
traffic of
every cell
Virtual traffic
of every cell
Quantity of
virtual channels
in every cell
Qu
antity
B o
f chan
nels av
ailably
pro
vid
ed b
y th
e cell
End
A<B
Yes
Downlink Capacity Estimation
c
aCCapacity ii )(
Quantity of equivalent voice channels:
7 10.75 1 76
Quantity of virtual channels in every cell
Look up the Erl B table and obtain that
the quantity of virtual channels required
by 3.07 Erl virtual traffic is 7
Quantity of equivalent voice channels to
be provided by every cellQuantity A of
channels to be
provided by the cell
Average
traffic of
every cell
Virtual traffic
of every cell
Quantity of
virtual channels
in every cell
Determine the number
of stations
Qu
antity
B o
f chan
nels av
ailably
pro
vid
ed b
y th
e cell
End
A<B
Yes
Downlink Capacity Estimation
Calculate the quantity of channels availably provided
by every cell based on power
])1[(/
)/(*1
/
)/(***
1
1
jj
N
j j
jj
N
j j
jjN
RW
NoEbv
RW
NoEbvLP
P
P represents the maximum service transmission power, which is 13 W
represents the noise power spectrum density on the front of the mobile
station receiver, and its value is -169 dBm
L represents the average path loss, which is evaluated by subtracting
6 dBm from the maximum path loss
j represents orthogoal factor, which is 0.6 for the multipath channel
represents interference factor from an adjacent cell. It is 0.65 for the three-
sector antenna macro cellj
Obtain that the quantity of equivalent voice channels actually provided by every cell is 71
NP
Quantity A of
channels to be
provided by the
cell
Average
traffic of
every cell
Virtual traffic
of every cell
Quantity of
virtual channels
in every cell
Determine the number
of stations
Quant
ity B
of
chann
els
availa
bly
provid
ed by
the
cell
End
A<B
Yes
Downlink Capacity Estimation
Quantity A of
channels to be
provided by the
cell
Average
traffic of
every cell
Virtual traffic
of every cell
Quantity of
virtual channels
in every cell
Determine the number
of stations
Qu
antity
B o
f chan
nels av
ailably
pro
vid
ed b
y th
e cell
End
A<B
Yes
Downlink Capacity Estimation
Comparison
The quantity of channels
to be provided by every
cell is 76
The quantity of channels
actually provided by every
cell is 71
There are 84 base stations
currently, and it cannot
satisfy downlink capacity
requirement, and some
stations should be added.
726588
717287
717286
717685
707684
697683
Number of
channels provided
Number of
channels required
Number of base
stations
488.095.1/88/8.40 Km
Downlink Capacity Estimation
Iterative calculation
If there are 88 base stations, the uplink and downlink coverage
capacity requirement can be met
In the case, the base station coverage radius is