Beruflich Dokumente
Kultur Dokumente
Confidentiality:
Internal
Dec 2010
Contents
1 PRINCIPLE................................................................................................ 4
1.1 High Speech Uplink Packet Access (HSUPA)............................................................ 4
2 TEST REPORT..................................................................................... 10
2.1 Simulation Result from HQ Test Lab............................................................................ 10
3 CASE STUDY......................................................................................... 19
3.1 Problem Description.................................................................................................... 19
Abstract:
This technical paper aims to give the readers the understanding about the relationship
of the HSUPA service and the uplink interference (3GPP terminology is RTWP,
The basic principle, test reports and case study from live network are included here to
give a clear view on how the HSUPA service impact on the uplink performance and how
Since HSPA service is widely used in the market, with a good understanding of this
topic, the engineer will be able to operate and optimize the network well.
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
1 PRINCIPLE
HSUPA is introduced within release 6 of the 3GPP specifications. It allows increased individual
connection throughputs, increased total cell throughputs and reduced round trip times.
In Huawei system (RAN10.0), the HSUPA data channel (E-DPDCH) can support either a 2 or 10
ms TTI.
- HSUPA Phase 1, 10ms TTI offers the benefit of improved physical layer performance
- HSUPA (Phase 2), 2ms TTI offers the benefit of reduced system delays and higher
Release 7 of 3GPP specification introduces the possibility of using 4 level Pulse Amplitude
Modulation (4PAM) which equivalent to 16QAM to increase the maximum achievable throughput,
In the WCDMA system, all the cells share the same frequency, which is beneficial to improve the
system capacity. However, co-frequency multiplexing causes interference among the users.
The Received Total Wideband Power (RTWP) on the uplink is the sum of all user signals and the
channel noise.
Noise Floor include of thermal noise and Noise Figure (NF) of the system.
PN K=
10Boltzmann *W ) + 1.38
log( K * Tconstant, NF
10 23 J / K
T Kelvin temperature, normal temperature: 290 K
If no other external factors e.g. external interference (from illegal sources), hardware issue. The
main factor that impact on Uplink interference is Traffic Load which includes traffic load from the
The relationship between uplink load and uplink interference (Noise Rise) is as below,
( = Load factor)
ITOT 1 1
NoiseRise = = =
PN N
1 UL
1 ( 1 + i ) Lj
1
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
UL load is affecting the noise level at the Node B receiver (Noise Rise).
50% is a good compromise between the number of sites and the offered capacity.
Too high uplink noise level cause cell shrink (reduction of coverage), breathing effect.
The relationship between Noise Rise and RTWP (Received Total Wideband Power),
RTWP (Received Total Wideband Power) dB = Noise Floor (Pn) + Noise Rise (Iown + Iother)
HSUPA service allows the users to achieve higher throughput on uplink but in the same time also
increase significant interference which causing the reduction of cell user access number and
It is great important to balance between the achievable throughput and system performance on
the uplink. Therefore, Huawei systems apply the Load Control mechanisms to control the level of
(More detail about Load Control Algorithms, please refer to Load Control Feature Document)
Load Control algorithms are different between HSUPA and R99 UL as follows:
1. Call Admission Control (CAC):
For HSUPA, PBR-based decision is used to check whether the QoS requirement of
existing users is fulfilled. The QoS is measured on the basis of the Provided Bit Rate
(PBR) of the users. If the QoS requirement is fulfilled, new users are allowed to access
the network.
-UL R99 is controlled by UL LDR trigger threshold if UL UU LDR algorithm is switch on.
-HSUPA, its scheduling is controlled by Maximum Target Uplink Load Factor and the
To ensure that the HSUPA user can access the cell, the minimum GBR (Guarantee Bit
Rate) is recommended. When set the GBR need to consider the trade-off between user
High GBR High user throughputHigh Uplink Interference Lower cell user number
Low GBR Low user throughput Low Uplink Interference Higher cell user number
(1)
2ms HSUPA
(2)
10ms HSUPA
(1)
2ms HSUPA
(2)
10ms HSUPA
To sum up, with different HSUPA TTI, 2ms and 10ms , the minimum guaranteed throughput of
user is vary 160kbps vs 64kpbs / 160kbps vs 32kbps. So, when the uplink load is limited, the
number of user that can access the cell is different (HSUPA TTI=2ms serve less number of user
per cell due to higher guaranteed throughput which generating higher uplink interference).
2 TEST REPORT
1RX_Antenna
UENum
ber Upl oadofAnt Throughput
enna% kbps
CAT5 8 74. 43 1215. 63
CAT6 8 86. 07 1163. 18
FTP user number (simultaneously upload) is limited at 20 due to the limited capacity of air
interface (uplink interference).
1RX_Antenna
UENum
ber Upl oadofAnt Throughput
enna% kbps
CAT5 20 76. 50 1100. 00
Cell throughput and cell user number are significantly improved especially on HSUPA
change the retransmission rate upon the Cell Load and UE Tx Power.
When cell load and UE Transmit Power are limited, the retransmission rate will be
When cell load and UE Transmit Power are less, the retransmission rate will be
decreased to completely utilize the resources and increase the effective rate of
UE.
Retransmission UE Transmit Eb/N0 * Coverage * UL UE peak Cell Throughput Max User
Power * Interference* throughput (Multi-users) Number
Large low low large low peak
low high more
Small high high small high high low less
(Please refer to the comparisons charts of Large and Small Retransmission performance in
next pages).
Test Case 1: HSUPA TTI=10ms 1Rx Antenna without Rx-Diversity (Indoor Case)
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
Test Case 2: HSUPA TTI=2ms 1Rx Antenna without Rx-Diversity (Indoor Case)
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
Test Case 3: HSUPA TTI=10ms 2 RXs Antenna with Rx-Diversity (Outdoor Case)
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
Test Case 4: HSUPA TTI=2ms 2 RXs Antenna with Rx-Diversity (Outdoor Case)
4000 120
3500
100
3000
Cell throughput
80
2500
Cell Load
2000 60
1500
40
1000
20
500
0 0
1 2 3 4 5 10 15 20 25 30 35 40 45 50 51 52 53
User Nmuber
4500 120
4000
100
3500
Cell Throughput
3000 80
Cell Load
2500
60
2000
1500 40
1000
20
500
0 0
1 2 3 4 5 10 15 16 17 20 25 29
User Number
Cel l Thr oughput Lar ge Ret r ansmi ssi on Cel l Thr oughput Smal l Ret r ansmi ssi on
Cel Load Smal l Ret r ansmi ssi on Cel l Load Lar ge Ret r ansmi ssi on
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
-In light cell load, small retransmission can give users higher throughput than large
retransmission.
-In medium and heavy cell load, large retransmission can achieve much more user
numbers and cell throughput than small retransmission; the gain can be 20% ~30%. (Based
on user numbers)
To achieve a maximum balance between cell capacity and peak rate of single user in uplink
(2ms TTI bring the higher peak rate of single user, while 10ms TTI can bring the
higher cell capacity). The operator can deploy another new feature (RAN12.0), HSUPA
TTI Auto Reconfiguration together with Adaptive Retransmission.
If the following condition has been fulfilled, this UEs TTI switch from 2ms to
10ms:
The Uu Load on Congestion Status & the UEs bit rate < Rate threshold for
2ms to 10ms
If both of the following conditions are fulfilled, this UEs TTI switch from 10ms to
2ms:
the UEs bit rate > Rate threshold for 10ms to 2ms.
UE Power is not limited. (If the 6A1 have been reported, the 6B2 shall be
reported after that.)
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
3 CASE STUDY
Based on the customers feedback in 3G Network (RAN10.0) of Operator XXX, the customers
had the difficulties to access the network ,sometime can access network but the data throughput
is very low and the connections frequently dropped. This problem is appeared at operators office
Symptoms:
-Based on the Statistic, we observed same pattern of high RTWP and PS drop in all cells that
circled in red.
-These cells are located in the same areas, nearby operators office.
-The highest RTWP appeared in Indoor cell, CHAMCHURI_C35-1 and the level of the uplink
interference of other outdoor cells reduce upon the distant from this indoor cell.
- Uplink interference (RTWP) increase very high reached the maximum value at -55 dBm during
-We also conducted the FTP download and upload at the Indoor cell, both give a very low
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
throughput.
Operators Office
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
Initially, we suspect that the uplink interference causing from the external source which may
illegally use in the operators building. This assumption was based on the interference pattern
In the first place, we did not expected that the issue related to traffic load due to based on
statistic, found that at the same number of HSUPA users, the RTWP didnt always high.
(Maximum =20 HSUPA users per cell, the HSUPA user number is limited by Maximum HSUPA
(1)
Use Spectrum Analyzer to search for External Interference source in operators building
(2)
Conduct Field Test (FTP upload) and open RNC LMT online trace measurement in the
Indoor Cell
Results
(1)
(2)
Based on the field test (FTP upload) and RNC LMT online measurement of indoor cell, we
HSUPA VS Uplink Interference Document
Confidentiality:
Internal
detected huge increment of RTWP when R99 test user doing the FTP upload (the test
user only can get R99 service due to HSUPA user number already hit maximum 20
Test Steps
(1) Condition before start FTP upload test, based on RNC LMT online measurement of
indoor cell, the HSUPA user was always at 20 (HSDPA user number was more
(2) We started FTP Upload on R99 (from Genex Probe, observed the average
throughput was around 300kbps) and the sudden increment of RTWP upto -55
dBm was observed via online measurement. (We randomly checked the RTWP of
(3) Next, we started FTP Upload on HSUPA (from Genex Probe, observed the very
low throughput, most of the time was 0 kbps) during that time observed no
(4) We checked the Load Control parameters setting of this indoor cell; found that the
LDR (Load Reshuffling) didnt turn-on. This is the reason why RNC still
scheduled bit rate to R99 users although current uplink load was high (above -95
dBm).
(5) For HSUPA user, the scheduling is based on Maximum Target Uplink Load
Factor which set to 75%, thus the bit rate was not scheduled to HSUPA user
Genex Probe
(6) We recommended to turned-on LDR (UL: BE RATE REDUCTION) for Indoor Cell
and monitor. Based on statistic after turn-on LDR, there was some improvement on
RTWP and PS drop but not on data throughput. However, high RTWP still
(7) We suspect there might be some issue with HSUPA service as well (We
implemented the HSUPA Phase2 (TTI 2ms) in this network) .Thus, we conducted
the HSUPA Load Test, to check how many HSUPA user (simultaneous FTP
From the test result, shown that for indoor case (with 1 Rx antenna), only
maximum 7 simultaneous upload users can supported. If the HSUPA user number
is exceed 7, will cause over-high RTWP. (Please refer to test result in previous
session)
of HSUPA Phase 2 with 1Rx antenna is limited and only can be improved
(1)
To turn-off the HSUPA phase 2 in current network and implement it later when upgraded
from RAN10.0 to RAN12.0 with Interference Cancellation and Dynamic NHR features
enabled.
(2)
To upgraded capacity of indoor site by adding 2nd carrier. This is to improve user experience
(3)
To turn-on Load Control (LDR) to control the interference level which generated from R99
users.
HSUPA VS Uplink Interference Document
Confidentiality:
Internal