Atoll Parameter Setting

Guideline & Sensitivity Analysis

for LTE FDD
80-W2838-4 Rev A
Oct 2010
Qualcomm Confidential and Proprietary
Restricted Distribution. Not to be distributed to anyone
who is not an employee of either Qualcomm or a
subsidiary of Qualcomm without the express approval of
Qualcomm’s Configuration Management.
 QUALCOMM Incorporated
5775 Morehouse Drive
San Diego, CA 92121-1714
U.S.A.

Copyright © 2010 QUALCOMM Incorporated.

All rights reserved.

Restricted Distribution. Not to be distributed to anyone who is not an employee of either Qualcomm or a subsidiary of
Qualcomm without the express approval of Qualcomm’s Configuration Management.
Not to be used, copied, reproduced in whole or in part, nor its contents revealed in any manner to others without the
express written permission of Qualcomm.
QUALCOMM is a registered trademarks of QUALCOMM Incorporated in the United States and may be registered in
other countries. Other product and brand names may be trademarks or registered trademarks of their respective
owners.
Nothing in these materials is an offer to sell any of the components or devices referenced herein. Certain
components for use in the U.S. are available only through licensed suppliers. Some components are not available for
use in the U.S.
This technical data may be subject to U.S. and international export, re-export or transfer (“export”) laws. Diversion
contrary to U.S. and international law is strictly prohibited.

Atoll Parameter Setting Guideline & Sensitivity Analysis for LTE FDD
80-W2838-4 Rev A
October 2010

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 2
 Table of Contents

 Executive Summary  Receiver Diversity Setting

 Transmitter Table  Inter Cell Interference

Coordination
 Cell Table
 Traffic Map Creation
 Performance Curve
 Clutter Class Properties
 Scheduler & Radio
Resource Management  Prediction Studies
 Traffic & Activity factor  Monte Carlo Simulation
Studies
 Throughput Demand
 Neighbor List Creation
 Terminal Property
 PCI Creation
 Uplink Analysis
 Summary – Parameter List

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 3
 Executive Summary

 The purpose of this document is to

 Provide an overview of the LTE FDD parameters used in Atoll
 Recommend the set of values and perform sensitivity analysis to ascertain their
impact on outcomes
 Use this as a future reference for LTE design activities

 Following Atoll software versions have been used for this purpose
 Atoll 2.8.2 Build 3357

 Conclusion
 Most of the parameters were found to have very high impact on the LTE FDD
performance metrics like throughput, SINR, RSRQ etc

 Some of these parameters significantly depend on actual vendor implementation and

should be used as per their recommendation

 For others, default or as recommended values may be used

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 4
Transmitter Table
 Transmitter Properties (1/2)

 Transmitter Properties

 The description & recommended settings are provided in the following slide

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 6
 Transmitter Properties (2/2)

Parameter Description Recommended Value Sensitivity

Default Cyclic Prefix Technique used by LTE to 0 – Normal Prefix (unless High
counter inter-symbol decided otherwise)
interference
PDCCH Overhead Can take upto 3 symbols in 4 symbols (BW<=3MHz) High
one subframe (includes 3 symbol (BW=5,10 MHz)
PCFICH, PHICH and PCH) 2 symbols (BW=15, 20MHz)

PUCCH Overhead Number of RB for PUCCH 2 RBs (BW=5MHz) High

4 RBs (BW > 5MHz)
Reference Signal Method of Calculation 0-Automatically Calculated Low
EPRE

Uplink Power Margin added to bearer 1dB (This should be set in Medium
Adjustment Margin selection threshold for accordance with the UL
safety against fast fading curve. If the UL curve is for a
given channel model, then
0dB should be used. If the
UL curve is set for e.g.
AWGN, then such margin is
necessary.
Adaptive MIMO Based on RS C/N or C/N Low
Switching Criterion C/(I+N)
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 7
 Transmitter Type

 Transmitter Type

 Two options are available here

 Intra-Network: With this setting, a
transmitter is treated both as a server
and an interferer
 Inter-Network : With this setting, a
transmitter is treated as an interferer
only (e.g. Femto HeNB, cross border
coordination,…)
 Recommended Setting : Intra-
Network
 Sensitivity of the Parameter: High.
As evident from the description, this
parameter impacts best server
determination as well as interference
analysis for prediction & simulation
studies and the results are highly
sensitive to this parameter.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 8
 Case Study – Impact of Tx Type

RSRQ (with Interferer) RSRQ (no Interferer)

 Study was performed on a

cluster of ~140 sectors.
Highlighted sites were used as
interferer (inter-sector)
 This has significantly impacted
overall RSRQ
 Similarly, PDSCH SINR has
also been impacted resulting in
throughput degradation

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 9
 Number of Antenna Ports

 Number of Antenna Ports

 These parameters are used to
determine MIMO / Diversity gains
 Transmission Antenna Ports: For
2x2 MIMO, this parameter shall be
set to 2
 Receiver Antenna Ports : This
parameter shall also be set to 2
 Sensitivity of the Parameter: High.
These parameters will significantly
impact LTE-specific analyses like
PDSCH SINR or throughput
calculation, use of SU-MIMO or
transmit diversity gain. The results
are therefore highly sensitive to these
parameters.
 While Tx antenna port setting will
impact DL performance, Rx antenna
port will change the UL performance

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 10
 Case Study – Impact of # Antenna

 Study was performed on a cluster of

~140 sectors. All parameters except
number of antennas were same
 With 2-ant, SINR has deteriorated
due to increased interference. This
has adversely impacted throughput
at lower SINR (where SU-MIMO gain
is not significant)
 At better SINR, higher SU-MIMO gain
compensated for the degradation
resulting in better throughput
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 11
 Allowable Frequency Bands

 This table is used to define the list of allowable frequency bands of operation, any one of which
can be selected on per cell basis in the cell table.
 Sensitivity of the Parameter: High. The frequency band and bandwidth of operation will impact
RF propagation as well as RSRP (EPRE) and hence the coverage. In addition, adjacent channel
suppression factor will impact the analysis of adjacent channel interference.
 Correct setting of frequency band of operation and associated parameter shown in the table
above is necessary to perform accurate simulation & prediction studies in Atoll.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 12
 TMA Equipment

 This table is used to define the TMA equipment that is going to be used at the sites.
 Sensitivity of the Parameter: High. Use of TMA equipment will significantly impact the
uplink performance including coverage and throughput estimation.
 Accurate settings of TMA gain, noise figure and transmission loss are necessary to ensure
realistic estimation of network performance.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 13
 Feeder Equipment

 This table is used to define the feeder equipment that is going to be used at the sites.
 Sensitivity of the Parameter: High. Transmission loss including feeder will significantly
impact the EIRP and hence the coverage.
 Accurate settings of these parameters are necessary to ensure realistic estimation of network
performance.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 14
 BTS Equipment

 This table is used to define the eNodeB noise figure and any additional losses that are taken
into account for link calculations.
 Sensitivity of the Parameter: Medium. Accurate settings of these parameters are necessary
to ensure realistic estimation of network performance.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 15
 LTE Equipment
Bearer Selection Quality Graph

MIMO Gain Graph

 Three key curves are set up under LTE equipment,
separately for downlink and uplink.
 Bearer selection: Selects best bearer
 Quality graph: Used to calculate effective throughput
 MIMO Gain: Represents MIMO gain

 These curves are critical to LTE simulations and

predictions and have been discussed in details in the
performance curve section

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 16
 LTE Bearer
LTE Bearer Table

 Provides a correspondence between radio bearer index and bearer efficiency.

 From the bearer selection graph (previous slide), Atoll first determines the best
supported bearer against a specific SINR.
 Then, LTE bearer table is looked up to determine the corresponding bearer efficiency
and used to calculate throughput.
 Bearer efficiency = #bit per modulation symbol x coding rate
 Sensitivity of the Parameter: High.
 Impact analysis of this table is performed in performance curve section.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 17
 Quality Indicator

Quality Indicator

 Represents the metric that should be used for calculation.

 Recommender setting; BLER
 Sensitivity of the Parameter: Low.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 18
 Scheduler Setting
Scheduler Table

 Atoll supports 4 types of scheduler viz. proportional fair, proportional demand, max, C/I and
round robin.
 Sensitivity of the Parameters: High.
 Recommended setting
 Scheduling method: As per infra. vendor implementation (prop. fair as default)
 Target throughput for voice/data services: 0-Peak rate
 Bearer selection criterion: 0-Bearer index
 Uplink BW allocation target: 2-Best bearer
 Impact of the above have been discussed in details in scheduler & radio resource
management and uplink analysis section
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 19
Cell Table
 Freq. Band & Channel Number

 Frequency Band & Channel Number  Frequency Band: The cell’s frequency
band from the Frequency Band list
 Channel Number: The number of the
channel from the list of available
channel number
 Recommended Setting: These
parameters should be set per market
basis, based on their spectrum and
bandwidth allocation
 Channel Allocation Status: could be
any because AFP is not implemented
 Sensitivity of the Parameters: High.
Definition of proper frequency band
(and hence bandwidth) is of utmost
importance as this will determine the
number of available resource blocks.
Channel number on the other hand, will
help perform external interference
analysis
 As channel BW increases, RS EPRE
for same max. PA will decrease
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 21
 Maximum Power

 Maximum Power
 Max Power (dBm): This field defines
the maximum transmit power per
antenna and is used to calculate
EPRE for different channels. If
however, “Reference Signal EPRE” is
set to “User-Defined”, this parameter
is not used
 Recommended Setting: This should
be set as per infrastructure vendors’
equipment capability. Assuming 30W
PA, this should be set to 44.8dBm
 Sensitivity of the Parameters:
High. The value of max. power will
directly impact the EPRE of different
LTE channels.
 For example, if max. power is
increased from 43dBm to 46dBm, RS
EPRE will increase by same amount
(3dB) and impact all prediction and
simulation analyses

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 22
 SS, PBCH, PDSCH, PDCCH Offset

 SS, PBCH, PDSCH, PDCCH EPRE Offsets w.r.to RS

 SS,PBCH,PDSCH,PDCCH Offset:
These fields define the difference in
energy of a resource element
belonging to SS / PBCH / PDSCH /
PDCCH with respect to energy of a
RS resource element.
 Recommended Setting: Unless
otherwise specified by the vendors,
these parameters should be set to 0.
 Sensitivity of the Parameters:
High. These values will determine
the EPRE of SS / PBCH / PDSCH /
PDCCH. These will also impact RS
EPRE and hence all subsequent
analyses and their outcomes
 Next slide shows the impact of offset
on RS EPRE

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 23
 Impact of Offsets on RS

 Impact of Offsets on RS EPRE

 To ascertain the sensitivity, RS EPRE was recalculated with change on offset

for one channel at a time
 Minimum impact of SS and PBCH was observed on RS EPRE calculation.
Minor impact was experienced due to PDCCH. PDSCH, on the other hand,
caused most severe impact

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 24
 Minimum RSRP Threshold

 Minimum RSRP: This field defines

the minimum RSRP required for a
user to be connected to the cell. If the
best server RSRP for a user is lower
than min. RSRP threshold, the user
will be rejected due to “no coverage”
 Recommended Setting: See table
 Sensitivity of the Parameters:
Medium. This value will impact the
percentage of users rejected due to
no coverage and hence all
subsequent analyses as well.
 Next slide shows the impact of Min.
RSRP on “no coverage”

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 25
 Impact of Min. RSRP Threshold

 Impact of Minimum RSRP on No Coverage (MC Simulation)

Min. RSRP: -113dBm Min. RSRP: -108dBm
Min. % of Users
RSRP Rejected due to
Target No Coverage
-113dBm 1.7%
-108dBm 4.7%
-103dBm 7.7%
-98dBm 14.3%

Min. RSRP: -103dBm Min. RSRP: -98dBm

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 26
 AMS Threshold (DL) (1/2)
DL Diversity Type AMS Threshold (dB)

 AMS & MU-MIMO Threshold (dB): This threshold is used in conjunction with diversity
support (DL) parameter, only when diversity support (DL) is set to AMS.
 In that case, AMS threshold represents the RSRQ threshold at which switching
between SU-MIMO and transmit diversity will happen.
 Recommended Setting for AMS: Diversity support (DL) should be set to SU-MIMO so
that AMS threshold is not used. Otherwise (with diversity support set to AMS), this
parameter threshold should be set to lower value (-19dB) so that in all cases, values
provided in the MIMO gain curves are used.
 Sensitivity of the Parameters: Low. This value will marginally impact throughputs

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 27
 Case Study – AMS Threshold (DL)

 Despite significant variation of AMS threshold between -4dB and -15dB, there was virtually
no impact on DL user throughput.
 This is due to the fact that at lower RSRQ (and hence PDSCH SINR), MIMO gain is
insignificant and hence did not impact the throughput results.
 Transmit diversity mode however showed inferior results as MIMO gain was never used

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 28
 UL & DL Traffic Load

 UL & DL Traffic Load

 Traffic Load (DL / UL): These fields
are used for prediction studies and
can also be imported from MC
simulation results
 Recommended Setting: To
represent the worst case predictions,
these values should be set to 100%
 Sensitivity of the Parameters:
High. Traffic load percentage will
significantly impact the load
dependent metrics like RSRQ,
PDSCH SINR, PUSCH SINR etc
 As example shown here as to how
traffic load has resulted in
deterioration of RSRQ statistics

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 29
 Max. UL & DL Traffic Load

 UL & DL Maximum Traffic Load

 Max. Traffic Load (DL / UL): These
fields are used to set the max. limits
beyond which cell will start rejecting
users
 Recommended Setting: Unless
otherwise recommended by the
vendors, these values should be set
to 100%
 Sensitivity of the Parameters:
High. These admission control
thresholds will significantly impact the
user performance
 Next slide shows an example of how
max. DL traffic load will impact the
user rejection due to resource
saturation
 These parameters can also be set in
the simulation window that overrides
cell table’s setting

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 30
 Impact of Max. DL Traffic Load

 Impact of DL Maximum Traffic Load (MC Simulation)

Max. DL Load = 50% Max. DL Load = 75%

Max. DL Load = 100%

Max. DL User Rejected due
Load to Resource
(%) Saturation

50 19.7%

75 11.0%

100 8.7%

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 31
 UL Noise Rise

Parameter Description Recommended Sensitivity

Value
UL Noise Rise (dB) Can be user-defined or an 0 - 1dB for unloaded High
output of Monte Carlo prediction;
simulations. This is the cell 3 - 6dB for loaded
specific value of uplink noise prediction
rise used for prediction studies (depending on load)

 As UL noise rise has gone up, UE

transmit power has also
increased.
 This increase in UE transmit
power was required to overcome
increased interference level at the
eNodeB receiver.
 Noise rise also impacts the uplink
coverage (percent area covered
within max UE transmit power).

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 32
 UL/DL Inter-technology Noise Rise

Parameter Description Recommended Sensitivity

Value
Inter-technology UL This noise rise represents the 0 dB High
Noise Rise interference created by mobiles and
base stations of an external network
on this cell on the uplink. This noise
rise will be taken into account in all
uplink interference-
based calculations involving this cell
in the simulation. It is not used in
predictions where Atoll calculates
the uplink total interference from the
uplink noise rise which includes
inter-technology uplink interference.
Inter-technology DL This noise rise represents the 0 dB High
Noise Rise interference created by mobiles of
an external network on the mobiles
served by this cell on the downlink.
This noise rise will be taken into
account in all downlink
interference-based calculations
involving this cell.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 33
 Max. Number of Users

 Max. Number of Users: This field defines the maximum number of simultaneously
connected users supported by the cell.
 Recommended Setting: Ideally, this value should come from the vendors. Till then,
this should be set to 2,000 so that scheduler saturation never happens
 Sensitivity of the Parameters: High. This parameter will decide the number of users
rejected due to scheduler saturation

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 34
 Impact of Max. Number of User

 Impact of Maximum Number of User (MC Simulation)

Max. # User= 10 Max. # User= 20

Max. # User= 30
Max. User Rejected due
Number of to Scheduler
Users Saturation

10 22.9%

20 1.7%

30 0.0%

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 35
 Fractional Power Control Factor (1/2)

 Fractional Power Control Factor Fractional Power Control Factor

 This factor is used for path loss
compensation when performing
fractional power control on the
uplink.
 For example, if this factor is set to
0.8, only 80% of the actual path
loss will be considered when
estimating the received power.
 Therefore, the received power
from any mobile on the uplink will
be estimated to be higher than it
would actually be (using 100% of
 Recommended Setting: Unless otherwise the path loss), which will be
recommended by the vendors, this interpreted by the mobile as a
parameter should be set to 1.0 need to reduce its transmission
power.
 Sensitivity of the Parameters: High. This
value will impact PUSCH SNR and also UE  This factor represents the
transmit power influence of the serving cell on the
fractional power of any mobile and
 Next slides shows the impact of FPC Factor impacts the UE transmit power.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 36
 Impact of FPC Factor (2/2)

 No significant impact was observed between FPC factor 1.0 and 0.9
 However, at 0.7, both UE transmit power and PUSCH SINR varied significantly.
 Unless otherwise recommended by the infrastructure vendors, this factor should be set to 1.0

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 37
Performance Curve
 Performance Curve (1/3)

 Atoll uses three curves and one bearer index table to convert PDSCH SINR to
throughput

 Based on calculated PDSCH SINR, Atoll refers to the best bearer to SINR curve
to ascertain the max. attainable modulation-coding combination (bearer)

 Next, it refers to the bearer index table to find out corresponding bearer
efficiency (bits/symbol)

 It then translates the bearer efficiency into instantaneous channel RLC

throughput (assuming 100% resources are available)

 Then, based on RRM, it calculates max. RLC throughput (corresponding to

available resources)

 Next, based on quality graph (BLR to SINR), it converts max. RLC throughput
to effective RLC throughput

 Then, SU-MIMO (or transmit diversity) gains are applied to determine the final
throughput

 Finally, based on throughput scaling factor RLC throughput is converted to

application layer throughput
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 39
 Performance Curve (2/3)
Start

 For Atoll, throughputs are derived

Calculate PDSCH SINR from the performance curves
 Accuracy of performance curve is
Determine Bearer Index from “Best therefore absolutely necessary to
Bearer to SINR” Curve
ensure design outcomes match with
Calculate Bearer Efficiency (Bits /
reality
Symbol) from “Bearer Index” Table
 Recommended Setting: To maintain
parity with the infrastructure, the
Calculate Channel Throughput (Max.
Throughput with 100% Resource) performance curves recommended
by the respective vendors should be
Perform Radio Resource Management used for design
for Selected Scheduler Type
 As an alternative, default 3GPP
curves provided by Atoll can also be
Calculate Peak RLC Throughput
used.

User “Quality Graph” (SINR to BLER);

 Sensitivity: High. As performance
Calculate Effective RLC Throughput curves represent the performance of
the respective infrastructure for a
Use “Max. SU-MIMO Gain” Curve; specific channel condition, proper
Determine Final RLC Throughput and accurate use is necessary
Use Throughput Scaling Factor;  Next slides show a comparative
Determine Application Layer Thpt. study between the performance
curves offered by two vendors
Stop
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 40
 Performance Curve (3/3)

Atoll Bearer Selection Atoll Quality Graph

Atoll Bearer Index Atoll MIMO Gain Graph

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 41
 SISO Throughput Comparison

 For SINR up to +14dB, marginal difference is observed. Beyond 14dB, vendor-2 is

significantly higher
 Vendor-2 graph corresponds to 9% BLER as against 0% for vendor-1
 Throughput (bps/Hz) = Bearer Efficiency (bit/symbol) x 14 / 15 assuming normal CP

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 42
 Max. SU-MIMO Gain Comparison

 For SINR above -6dB, vendor-1 gain is always higher. The difference is significant
beyond +8dB
 Below -6dB, Vendor-1 gain is lower than 1. This needs to be reviewed to ensure
accurate Atoll operation

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 43
 SU-MIMO Throughput Comparison

 As a combined impact of the two previous curves, vendor-2 throughput above +5dB
was found to be significantly lower.
 As part of Atoll simulation result, this should result in higher user throughput with
vendor-1 curve

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 44
 Project Details
Clutter Traffic

 140 sectors considered for analysis

 The clutter consists of dense urban, residential and commercial industrial
 Traffic distribution varies between 8 to 25 FTP users per sectors (90% activity factor)
 MC simulation run for 2x2 MIMO with 10 simulations
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 45
 PDSCH SINR Distribution
(MC Users)
Vendor-1 Vendor-2

 As expected, no change in PDSCH SINR distribution was experienced

 Above results are based on MC user (13.4 users on the average per sector)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 46
 RLC User Throughput Distribution
(MC Users)

 Average peak RLC user throughput was found to be 20% higher with Vendor-
1 curves
 Same observation is drawn from the cells (CDF) throughput distribution too

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 47
 Peak Channel Throughput
Distribution (User Location)

 Instantaneous channel throughput was found to be higher with Vendor-1 curves

 For vendor-2, all users with PDSCH SINR > 21dB, were assigned bearer index 28
 For Vendor-1 however, granularity was provided upto SINR = 35dB

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 48
 Conclusion

 Comparative analyses of Vendor-1 & Vendor-2 graphs show

 For SINR upto +14dB, marginal difference is observed between the two SISO graphs.
Beyond 14dB however, Vendor-2 shows significantly higher throughput
 For SINR above -6dB, Vendor-1’s max. SU-MIMO gain is always higher. The
difference is significant beyond +8dB
 Below -6dB, Vendor-1’s max. SU-MIMO gain is lower than 1. This needs to be
reviewed to ensure accurate Atoll operation
 As a combined impact of the two curves, Vendor-2 throughput above +5dB was found
to be lower than Vendor-1

 Atoll Simulation
 Despite nearly identical PDSCH SINR distribution, average RLC user throughput is
20% higher in case of Vendor-1. This is due to better performance curve used
 Sensitivity
 High
 Recommendation
 In order to accurately represent the real infrastructure performance, it is necessary to
use the vendor-recommended set of curves for Atoll simulation
 If above curve is not not available, as an alternative, 3GPP curve provided as default
in Atoll can be used

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 49
Scheduler & Radio Resource
Management
 Scheduler Algorithm (1/2)

 In Atoll, following scheduling algorithms are used:

 Proportional Fair
 Goal: Distribute resources among users fairly in such a way that, on the
average, each gets the highest possible throughput under its radio condition
 This method allocates the same amount of resources to all the users with a
maximum throughput demand
 The resources allocated to each user are either the resources it requires to
achieve its maximum throughput demand or the total amount of resources
divided by the total number of users in the cell, which ever is smaller

 Proportional Demand
 Goal: Distribute channel throughput among users proportionally to the
demands
 This method allocates resources proportional to the demands of the users who
have a maximum throughput demand.
 Therefore, users with higher maximum throughput demands will have higher
resulting throughputs compared with the lower maximum throughput ones

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 51
 Scheduler Algorithm (2/2)

 Round Robin
 Goal: Allocate equal resources to users fairly

 This scheduling method allocates the same amount of resources to all the
users with a maximum throughput demand

 Therefore, the resources allocated to each user are either the resources it
requires to achieve its maximum throughput demand or the total amount of
resources divided by the total number of users in the cell, which ever is smaller.

 Max. C/I
 Goal: Achieve maximum aggregate cell throughput

 This scheduling method allocates the resources required by the users to

achieve their maximum throughput demands in the order of their PDSCH
C/(I+N) in downlink and of their PUSCH & PUCCH C/(I+N) in the uplink.

 This means that users who are under good radio conditions will get all the
resources they require

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 52
 Radio Resource Management (1/3)
Start

 The process is executed in

Determine Total DL Resource Per Cell
following steps

Determine Total UL Resource Per Cell  Based on traffic demand, the

scheduler selects a specific number
# User<=
of mobiles for scheduling and RRM
Y Retain All Users
Max. User? process
N
 This is followed by calculation of
Scheduler Saturation; Retain Max. minimum and maximum throughput
Users Defined in Cell Table
demands
Sort Users in Descending Order of  Next, resources are allocated to meet
Service Priority
minimum throughput demands
Allocate Resource to Meet Min. Thp  In case resources are still available,
Demand; Start with First User
throughput demands are addressed
based on scheduler algorithm
Resource
Remains?  Proportional Fair
N
Y
 Proportional Demand
Allocate Resource to Users with Max.
Thp Demand; Use Scheduling Algo.  Round Robin
 Max. C/I
Generate Results

Stop

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 53
 Radio Resource Management (2/3)

 As a first step, Atoll determines total resources for each cell (UL & DL)

 Then, scheduler checks whether the number of users per cell exceeds the
max. number of users defined. If yes, Atoll retains max. number of users and
rejects others for scheduler saturation

 Next, scheduler starts allocating resources to the users, one by one in order of
priority, to satisfy min. throughput demand

 If min. throughput demand cannot be met for all users, user rejection happens
(termed as resource saturation)

 Then, scheduler calculates

 Remaining resource at every cell (after satisfying minimum throughput demand) and
 Resource demand per user to meet their remaining throughput demand (max.
demand – min. demand)
 Resource Demand Per User = Remaining Throughput Demand of the User / Peak
Channel Throughput of the User

 Finally, based on scheduler algorithm (described in the next slide),

incremental resources are allocated to the users

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 54
 Radio Resource Management (3/3)

 Proportional Fair
 Resource Allocated to a User = Min. [Resource Demand of the User, Remaining
Resource at the Cell / Total Number of Users]
 Proportional Demand
 Resource Allocated to a User = Min. [Remaining Resource at the Cell, Sum of
Resource Demands of all Users] x [Resource Demand of the User / Sum of
Resource Demand of All Users]
 Round Robin
 Resource Allocated to a User = Min. [Resource Demand of the User, Remaining
Resource at the Cell / Total Number of Users]
 Max. C/I
 Resource Allocated to a User = [Throughput Demand of the User / Downlink
Peak Channel Throughput for the User]
 Resource allocation process continues until demands of all users are met or
max. cell load is reached
 The equations for proportional fair and round robin are identical except for the
fact that in case of proportions fair, multi-user gain (MUG) is added to
represent benefit from fast fading

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 55
 Case Study – Impact of Scheduler
(1/2)

 Case study was performed on a cluster of ~140 sectors. All parameters except
scheduler type were kept unchanged
 As expected, PDSCH SINR, being independent of scheduler type, remained same
 Peak channel throughput is also calculated without taking scheduler into account
(assumes 100% resource availability). As a result, all except proportional fair
scheduler showed nearly identical results
 The difference with proportional fair scheduler is attributable to MUG (multi-user
gain) that represents benefit of fast fading (discussed later)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 56
 Case Study – Impact of Scheduler
(2/2)

 Effective RLC user throughput is dependent on the type of service, throughput

demand as well as scheduler algorithm
 Above graph shows that despite similar PDSCH SINR, scheduler algorithms have
resulted in significant difference in end user throughput perception
 As expected, Max. C/I demonstrated highest throughput with proportional fair
being closest
 No impact was seen on user rejection or any other performance metrics
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 57
 Multi User Gain (MUG)

 Multi-user gain represents the benefit from fast fading in a multi-user environment;
as the number of simultaneous users in a cell increases, the gain also goes up

 This is used only for proportional fair scheduler to model the use of resource
scheduling over a period of time

 In Atoll, to represent its impact, a “multi-user gain” curve (next slide) is used along
with proportion fair scheduler definition

 Depending on the number of simultaneous users being served by a cell, peak

channel throughput (for any particular user at a specific PDSCH SINR) is
multiplied by the corresponding multi-user gain value

 As a result, both peak channel thpt (pre-scheduler) and effective RLC thpt
(post-scheduler) show better results compared with no-MUG scenario

 As a result, both peak channel t’put (pre-scheduler) and effective

RLC t’put (post-scheduler) show better results compared with no-
MUG scenario

 MUG in actual network are a results of the variability of the RF conditions,

and thus of the reported CQI
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 58
 Case Study – Impact of MUG (1/2)

 Case study was performed on a

cluster of ~140 sectors. All
parameters except MUG were kept
unchanged
 As expected, PDSCH SINR, being
independent of scheduler type,
remained same
 Peak channel throughput varied due
to change in MUG
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 59
 Case Study – Impact of MUG (2/2)

 Effective RLC user throughput is dependent on the type of service, throughput demand
as well as scheduler algorithm
 Left graph shows that despite similar PDSCH SINR and use of same scheduler, MUG has
significantly impacted end user throughput perception
 Except for MUG, as proportional fair scheduler does not have any difference with round
robin, with No-MUG, the two have shown nearly the same results (right graph)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 60
 Conclusion

 Even for the same PDSCH SINR distribution, effective RLC user
throughput will significantly vary with the choice of scheduler type

 In case of proportional fair scheduler, multi user gain (MUG) curve

significantly impacts both peak channel throughput and RLC user
throughput

 Sensitivity
 Scheduler Type : High
 Multi User Gain Curve (for Prop. Fair scheduler) : High

 Recommendation
 Atoll scheduler implementation does not consider the time variations (static
simulations)
 For a given type of scheduler (typically proportional fair or proportional demand),
estimated capacity (user and cell t’put) can only be consider in a relative way:
improvement in RF condition showing improvement in t’put
 MUG should be set to 1 (no gain)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 61
Traffic and Activity Factor
 User Generation & Distribution (1/2)

 In Atoll, users generation is impacted by the following:

 Traffic Data: This input is assigned per morphology weighted by Poisson’s distribution
 User Profile: Each UE is assigned a specific service, mobility type and terminal according
to the user profile assigned to it
 Activity Status: A transmission status is determined according to the activity probabilities
(DL-only, UL+DL or UL-only).
 Indoor/outdoor User: Determined based on indoor user percentage defined per
morphology
 Geographical Location: Geographical location is determined randomly; the number of
users per morphology is determined by traffic data

 Activity factor impacts the number of users generated by Atoll (in the
example DL user ratio matches closely with DL activity factor)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 63
 User Generation & Distribution (2/2)

 Following are the corresponding sample Atoll setup

Traffic Map Service Definition

Traffic Properties Terminal Type

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 64
 Case Study – Impact of Activity Factor

 Case study was performed on a cluster

of ~140 sectors. All parameters except
DL activity factor were kept unchanged
 Although SINR or RSRQ did not change
due to nearly 100% DL load, average
user throughput has increased due to
lower activity
 If activity factor becomes so low that
max. demand is satisfied at lower DL
load, RSRQ & SINR will show change

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 65
 Conclusion

 Even for the same RSRQ & PDSCH SINR distribution, effective RLC
user throughput will significantly vary with change in activity factor
 This is due to the fact that cells could use same number of downlink
resource blocks to support lower total throughput demand
 In this example however, max. throughput demand of the users were
not met. As a result, cell load continued to remain unchanged (and
hence RSRQ and SINR)
 If however, had max. throughput demand been achieved, cell load
would have gone down improving both RSRQ and SINR
 Sensitivity : High
 Recommendation
Type of Activity Min. Throughput Max. Throughput
Comments
Service Factor (kbps) (kbps)

FTP 90% 0 6,000 a) The max. targets should be derived from the analysis of
existing 3G network
HTTP 10% 0 256 b) For non-GBR services, min. can be set to 0; for GBR
VoIP 2.5% 320 320 services however, min. throughput should be set as per
service requirement
Streaming 90% 0 512 c) Owing to tool’s limitation, only best effort type
application (FTP) is recommended for use at this stage

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 66
Throughput Demand
 Throughput Demand

 In Atoll, two sets of throughput demands are used

 As a first step, Atoll tries to satisfy all users of their minimum throughput demand.
 If minimum throughput demand cannot be met for all users with available resources, Atoll
rejects low priority users (called resource saturation)
 Otherwise, as a next step, Atoll determines incremental demand for all users as well as
available resources
 Then, based on scheduler algorithms, it allocates resource to the users to meet their
maximum demand
 This process continues until all resources are used up or all demands are met

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 68
 Case Study –
Impact of Min. Throughput Demand
Min. Thp Demand= 0kbps Min. Thp Demand= 300kbps Min. Thp Demand= 500kbps

 As minimum throughput target has

increased, resource saturation has
gradually gone up
Min. User
Thp. Rejected due
 With increasing minimum throughput
Demand to Resource demand, as scheduler started being
(kbps) Saturation
into operation later, this has
0 0.0% adversely impacted effective RLC
300 0.9% user throughput distribution
500 9.6%
 No accepted user was found to be
with thp lower than min. target (lower
section of the graph)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 69
 Case Study –
Impact of Max. Throughput Demand

 Case study was performed on a cluster of ~140 cells. All parameters except max.
throughput demand were kept unchanged
 Although the scheduler uses the max. throughput demand to calculate additional
resource requirement, this does not impact RF performance
 As a result, PDSCH SINR remains unchanged
 The max. throughput demand however partially impacts the RLC throughput distribution.
As for example, if max. throughput demand is set to 2Mbps, no user will ever assigned
resource for higher throughput

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 70
 Conclusion

 Both minimum and maximum throughput demand have some impact

on MC simulation results

 Minimum throughput demand impacts both resource saturation

statistics and user throughput distribution

 Maximum throughput demand, on the other hand, may slightly impact

average throughput

 Sensitivity
 Min. Throughput Demand : High
 Max. Throughput Demand : Low
 Recommendation
 Slide-65 provides details of service settings
 Ideally, these should be set in line with application usage and demand experienced
in the existing network.
 Also, this can be set to the maximum supported value of a particular device type

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 71
Terminal Property
 Terminal Property (1/2)

 Terminal Property

 Diversity Support: Considering the

availability of MIMO in the network,
this should be set to MIMO
 Number of Tx Antenna Ports : As
most likely, UE will have 1 transmit
antenna, this should be set to 1
 Number of Rx Antenna Ports: UE
will have 2 antennas – this should be
set to 2
 Sensitivity of the Parameter: High.
As evident from the description,
these parameters will impact MIMO
and Tx/Rx diversity performance
 In order to ensure accurate prediction
and simulation results, these should
be set as indicated above

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 102
 Terminal Properties (2/2)

Parameter UE – MIMO UE – MIMO UE – Non-

Smartphone Data Card MIMO

Min Power -40 -40 -40

Max Power 21 23 23

Gain (dBi) 0 0 0

Losses (dB) 5 5 5

Noise Figure (dB) 8 8 8

DL Rx DL Rx DL Rx
Reception Equipment
Equipment Equipment Equipment
Antenna

Diversity Support MIMO MIMO (none)

No. of MIMO Tx Ant 1 1 1

No. of MIMO Rx Ant 2 2 1

 Losses represents negative antenna gain, and thus should be set according to the
frequency band
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 103
 Key Parameters in an LTE Budget – UE
Antenna Gain

2
UE Antenna Gain [dBi]

-3

-8

-13

-18
0 500 1000 1500 2000 2500
Frequency [MHz]

Specified frequency operation band will determine the

expected UE antenna gain.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 104
 Case Study 1 –
Impact of Terminal on Downlink

 Study was performed on a cluster of

~140 sectors. All parameters except
terminal type were same
 As expected, PDSCH SINR had no
impact due to terminal type
 At higher SINR however, owing to
SU-MIMO gain, MIMO terminal has
demonstrated much better channel
and user throughput

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 105
 Case Study 2 –
Impact of Terminal on Uplink

 Study was performed on a cluster of ~140 sectors. All parameters except terminal
type were same
 Also, in this specific example, UL max. SU-MIMO gains were set to a unique value
(~0.82)
 As expected, PUSCH/PUCCH SINR had no impact due to terminal type

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 106
Uplink Analysis
 Uplink Calculation & Bearer
Determination (1/3)

 First, Atoll calculates the allowed maximum transmission power for

the terminal type used by the pixel, subscriber, or mobile. This power
is calculated by performing fractional power control.

 Next, Atoll calculates the uplink noise over the channel bandwidth
used by the cell.
 The used bandwidth depends on the number of used subcarriers.
 The uplink noise comprises thermal noise and the noise figure of the
equipment.

 Thereafter, the PUSCH and PUCCH interference are calculated for

Monte Carlo simulations
 If activated, interference reduction due to static uplink ICIC is also taken care of
for the cell edge users

 Finally, Atoll takes the ratio of the signal level and the sum of the total
interference & noise to determine C/(I+N)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 108
 Uplink Calculation & Bearer
Determination (2/3)

 In Atoll, best bearer index selection will be done as follows:

 The bearer whose index is less than or equal to the highest bearer index
defined for the service being accessed
 Whose selection thresholds are less than the PUCCH & PUSCH C/(I+N)
calculated
 If the cell supports Receiver Diversity or AMS, the receiver diversity gain, ,
corresponding to the bearer is applied to its selection threshold
 The additional uplink diversity gain defined for the clutter class of the pixel,
subscriber, or mobile is also applied.

 Once best bearer index is known, PUCCH/PUSCH SINR is calculated

as under:
 For Receive Diversity
 PUCCH/PUSCH SINR = Calculated PUCCH/PUSCH SINR + Rx Diversity
Gain + Addl. UL Diversity Gain
 For AMS
 PUCCH/PUSCH SINR = Calculated PUCCH/PUSCH SINR + Rx Diversity
Gain + Addl. UL Diversity Gain (where RSRQ < AMS Threshold)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 109
 Uplink Calculation & Bearer
Determination (3/3)
Bearer Table Atoll Cell Table

Bearer Selection Graph Clutter Class

Rx Diversity Gain

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 110
 Uplink Bandwidth Allocation (1/4)

 In Atoll, the uplink bandwidth allocation depends on the target

defined for the scheduler used by the cell

 Full Bandwidth
 Full channel width is used by each mobile in the uplink. As there is no reduction in
the bandwidth used for transmission, there is no gain in the PUSCH and PUCCH
C/(I+N)

 Issues with Full Bandwidth

 As full bandwidth is allocated to all users irrespective of their service needs or any
other considerations, this will turn out to be the most inefficient method.
 Lack of optimization of channel bandwidth for interference management will produce
highest uplink interference that might ultimately result in user rejection due to “no
service” in the uplink (i.e. PUSCH SINR < min. PUSCH SINR threshold)
 Full bandwidth allocation will happen even for services like VoIP that will never
require more than a few RBs.
 As a result, in all scenarios, this method will manifest worst performance.
 The impacts have been further discussed in the case study (provided later)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 111
 Uplink Bandwidth Allocation (2/4)

 Maintain Connection
 The bandwidth used for transmission by a mobile is reduced only if the PUSCH and
PUCCH C/(I+N) is not enough to access the lowest bearer.
 For example, as a mobile moves from good to bad radio conditions, the number of
frequency blocks used by it for transmission in uplink are reduced one by one in
order to improve the PUSCH and PUCCH C/(I+N)

 Issues with Maintain Connection

 Like full bandwidth, this method also starts with allocating full bandwidth to each
user, irrespective of their needs or any other considerations.
 Then, if the scheduler finds that PUSCH SINR experienced by a user is lower than
min. threshold (that would otherwise have resulted in rejection due to “no service”), it
gradually reduces RB allocation of that user by 1 RB at a time, until PUSCH SINR
becomes equal to or higher than the threshold.
 That way, UL user rejection due to “no service” is avoided by this method that tries to
ensure that all connections are maintained. But, it does not at all try to optimize the
bandwidth allocation by reducing it for users who do not need it.
 As, by default, full bandwidth is allocated to all users irrespective of their service
needs, this will turn out to be an inefficient method as well.
 The impacts have been further discussed in the case study (provided later)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 112
 Uplink Bandwidth Allocation (3/4)

 Best Bearer
 The bandwidth used for transmission by a mobile is reduced in order to improve the
PUSCH and PUCCH C/(I+N) enough to access the best bearer.
 For example, if using 5 frequency blocks, a mobile is able to access the best bearer,
and using 6 it would only get access to the second best, it will be assigned 5
frequency blocks as the used uplink bandwidth

 Issues with Best Bearer

 This approach, unlike the other two, tries to allocate RBs to a user such that these
correspond to the best bearer of that user’s RF condition.
 This method also completely ignores that fact whether a service at all needs that
many RBs and continue to unnecessarily allocate those in a sub-optimal manner.
For example, a VoIP user may be allocated 10 RBs whereas only 2-3 would have
been sufficient.
 Nevertheless, out of the three methods, this is found to be the most reasonable
technique, although it does not attempt to really optimize resource allocation based
on service demand.
 The impacts have been further discussed in the case study (provided later)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 113
 Uplink Bandwidth Allocation (4/4)

 UL BW Allocation: Three probable setting are provided viz. full BW, Maintain connection
and best bearer
 Recommended Setting: As per vendor’s implementation. Otherwise, as a default, Best
Bearer should be used
 Sensitivity of the Parameter: High. As evident from the description, this parameter will
impact uplink performance

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 114
 Case Study (1) – Impact on UL FTP
Clutter Traffic

 140 sectors considered for analysis.

 The clutter consists of dense urban, residential and commercial industrial.
 Distribution varies between 8 to 25 FTP users per sector (10% UL activity factor).
 MC simulation run for 2x2 MIMO with 10 simulations.
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 115
 Case Study (1) – Impact on UL FTP

No UL
Service

 Both “Full BW” and “Maintain Conn”

demonstrated almost similar
performance.
 With “Full BW”, ~10% of the users
experienced SINR < min. threshold of
-12dB. “Maintain Conn” however
managed to retain those users.
 Owing to its different BW allocation
approach, “Best Bearer” has shown
better SINR & UE Tx power.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 116
 Case Study (1) – Impact on UL FTP

UL Bandwidth UL User Rejection

 Full BW allocation to all UL users
caused significant UL interference
Allocation (No Service)
and user rejection due to no service.

Full Bandwidth 10.56%  Maintain connection however tried to

improve the situation and could bring
down the rejection to ~1.5%.
Maintain Connection 1.46%
 Best bearer allocated a maximum of
15 RBs to any user, causing
Best Bearer 0.00% minimum UL interference and hence
no user rejection.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 117
 Case Study (2) – Impact on UL VoIP
Clutter Traffic

 140 sectors considered for analysis.

 The clutter consists of dense urban, residential and commercial industrial.
 Distribution varies between 8 to 25 VoIP users per sector (60% UL activity factor).
 MC simulation run for 2x2 MIMO with 10 simulations.
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 118
 Case Study (2) – Impact on UL VoIP

 Even for a service like VoIP, where

more than a few RBs is never
needed, all three methods resulted in
significantly large RB allocation.
 This is attributable to the fact that all
BW allocation methods ignore
service type.
 Dependency of BW allocation on best
bearer resulted in slightly better SINR
for this method.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 119
 Conclusion

 Based on the case studies, following conclusions are made

 All three bandwidth allocation techniques completely ignore RB assignment
according to its requirement for any specific service type.
 None of the three methods tried to optimize UL interference by manipulating RB
allocation as per service requirement
 “Full bandwidth” method always allocates all RBs to all users. This causes significant
UL interference and user rejection due to poor SINR
 “Maintain connection” tries to retain users by reducing RB allocation of those users
that would have otherwise been rejected due to poor SINR. There is no attempt for
real UL interference optimization though.
 “Best bearer” performs better than other two; however, this also does not
attempt to allocate RB as per service requirement.
 Although full BW and maintain connection may be somewhat usable for FTP-only
scenario, these cannot be used for mixed service scenarios like FTP & VoIP.
 Even “best bearer” is not very suitable for mixed traffic scenario. As RB allocation
happens without paying attention to the service type, unnecessary allocation of
larger RB chunks will increase interference and result in poor performance.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 120
 Recommendation

 For UL, selection UE’s transmission BW is very challenging as this

will impact both throughput and battery life.
 The wider the BW allocation, the lower will be EPRE. This will, in turn, lower SINR
and hence throughput.
 To meet same demand, additional RBs will therefore be needed. This will result in
increased interference as well as lower battery life.

 Short-term recommendation
 Until new bandwidth allocation schemes are developed, use of “best bearer” method
will result in relatively less incorrect uplink design outcomes.
 For FTP service, maximum allowable uplink bearer can be set to the ones
recommended by the vendors.
 For VoIP service, however, assuming 12.2kbps throughput demand can be served
by 2 RBs, best bearer should be limited to such a value that will ensure 2 RBs with
corresponding bearer efficiency offers at least 12.2kbps throughput.
 .

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 121
Receiver Diversity Setting
 Receiver Diversity Setting
DL Diversity Type UL Diversity Type

 Diversity Support (DL / UL): These fields are used for to define the diversity support for DL
& UL directions. Depending on these settings, SU-MIMO gains are used for throughput
calculations
 Recommended Setting: These should be set to SU-MIMO for DL & Rx Diversity or MU-
MIMO (if supported) for UL
 Sensitivity of the Parameters: High. As these will determine the use of SU-MIMO / Rx
Diversity gains, throughput calculations will be significantly impacted by these

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 123
 Case Study – Impact of Downlink
Diversity Setting

 Study was performed on a cluster of

~140 sectors. All parameters except DL
diversity type were same
 As expected, no impact was observed
on PDSCH SINR
 Nevertheless, both peak channel
throughput and effective RLC user
throughput have been impacted due to
SU-MIMO gain (at higher rate)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 124
 Case Study – Impact of Uplink
Diversity Setting

 Study was performed on a cluster of ~140 sectors. All parameters except UL diversity
type were same
 As expected, no impact was observed on UL PUCCH/PUSCH SINR
 Nevertheless, uplink throughput has been significantly impacted (inferior results with SU-
MIMO); this is due to the fact that UL SU-MIMO gain was set to <1
 Recommended Setting: As per vendor’s implementation. Otherwise, as a default, Rx-
Diversity / MU-MIMO (if supported) should be used
 Sensitivity of the Parameter: High. As evident from the analysis, this parameter will
impact uplink performance
80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 125
Inter-Cell Interference
Coordination (ICIC)
 ICIC (1/3)

 What is ICIC?
 ICIC stands for inter-cell interference coordination.
 This is a technique used to minimize the inter-cell interference experienced by
the cell-edge users

 Types of ICIC
 ICIC can be of two types viz. static ICIC and dynamic ICIC.
 Static ICIC is a fixed fractional frequency allocation approach (1/3) to the cell-
edge users that does not change over time
 Dynamic allocation varies over time and is coordinated via X2 interface to
avoid collision over both frequency and time. Also, in dynamic allocation mode,
allocation is adapted over time to match traffic requirement

 What is current implementation in Atoll?

 Only static ICIC. Dynamic ICIC is not implemented

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 127
 ICIC (2/3)

 Which parameters model static ICIC in Atoll?

 Frequency band of operation and number of channels. These will determine
how much spectrum will be available at the cell edge users
 Definition of PCI (physical cell ID).
 This is a necessity as PCI will determine RB allocation to the cell-edge users. PCI
consists of PSS ID (0,1,2) and SSS ID (0,…167).
 In static ICIC mode, entire spectrum is divided into three equal chunks viz. fraction 0,
fraction 1 and fraction 2 where fraction 0 corresponds to PSS ID =0, fraction 1
corresponds to PSS ID=1 and fraction 2 corresponds to PSS ID=2.

 Interference coordination support:

 The type of inter-cell interference coordination technique supported by the cell (Static
DL and/or Static UL)

 ICIC delta pathloss threshold (dB):

 As ICIC applies to the cell-edge users only, there needs to be a mechanism for Atoll
to understand what is the definition of cell-edge user.
 This user-defined parameter is used for this purpose. This is the delta between best
server and second best server’s pathloss.
 If the delta pathloss for any pixel is found to be less than this value, the pixel and the
corresponding users will be treated as cell-edge and ICIC will come into play.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 128
 ICIC (3/3)

 Which parameters model static ICIC in Atoll?

 ICIC ratio (DL) (%):
 This is the percentage of the total downlink traffic load that corresponds to the ICIC
part of the frame.
 For example, if the downlink traffic load is 80%, and ICIC ratio is set to 50%, it means
that 40% of the downlink traffic load is on the ICIC part of the frame while the other
40% is on the non-ICIC part.
 This can be user-defined or can be derived from Monte Carlo simulations

 ICIC UL Noise Rise (dB):

 The uplink noise rise of the part of the LTE frame that uses a fraction of the channel
bandwidth.
 This noise rise is only used when the ICIC support for the cell includes static UL. This
can be user-defined or an output of Monte Carlo simulations.
 This is a technique used to minimize the inter-cell interference experienced by the
cell-edge users

 Sensitivity: High. ICIC will significantly impact analyses outcomes

 Recommendation : Should be used if supported by the vendors

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 129
 Case Study – ICIC Delta Path Loss (1/2)

 ICIC Delta Path Loss Threshold (dB)

 ICIC Delta Path Loss Threshold:
The maximum difference between the
path loss of the second best server
and the path loss of the best server to
be considered at cell edge. Cell edge
mobiles are served by the ICIC part
of the LTE frame, i.e., the part of the
frame that uses a fraction of the channel
bandwidth.
 Recommended Setting: This
parameter should be set if and only if
ICIC is supported by the network.
Otherwise this should be set to 0dB
 This parameter will work only if
interference coordination support is
checked
 Sensitivity of the Parameters: High.
This value will impact SINR and
Throughputs

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 130
 Case Study – ICIC Delta Path Loss (2/2)

 With increase in delta path loss threshold, ICIC has come into play for more users,
especially the ones who are located at the cell-edge or far from cell centre.
 As a result, the PDSCH SINR has improved. As ICIC is mainly aimed at cell-edge users,
the improvement has been more visible at the lower end of the PDSCH SINR curve.
 Owing to improvement in SINR, DL effective user throughput has also increased.
 Overall, high impact of ICIC has been observed on the simulation results.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 131
 Case Study – ICIC Ratio (DL)

 ICIC Ratio (DL) (%): This is the percentage of the total downlink traffic load that
corresponds to the ICIC part of the frame.
 As the ratio is changed from 0% to higher, ICIC has resulted in improvement in PDSCH
SINR distribution. Once ICIC kicked in, the impact however has been minimal.
 Recommended Setting: This parameter should be set if and only if ICIC is supported by the
network. Otherwise this field should be kept blank. This parameter will work only if
interference coordination support is checked.
 Sensitivity of the Parameters: High. This value will impact SINR and Throughputs

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 132
Traffic Map Creation
 Traffic Map Creation

 The first table defined the number of FTP users per sector. Depending on the service mix, this
table can also be used to defined the user number of other service types.
 The second table defines the weights of individual clutters that will determine number of users
generated per clutter classification.
 This table also defines the percentage of indoor users per clutter category that shall be used
during Monte-Carlo simulation studies.

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 134
Clutter Classes Properties
 Clutter Classes Properties

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 136
Prediction Studies
 Prediction Studies

Recommended List Comments

Downlink

Best RSRP (RS EPRE) Level (dBm) From “Effective Signal Analysis”
RSRQ Level (DL) (dB) From “Coverage by C/(I+N) Level (DL)”
PBCH C/(I+N) Level (DL) (dB) From “Coverage by C/(I+N) Level (DL)”
PDSCH C/(I+N) Level (DL) (dB) From “Coverage by C/(I+N) Level (DL)”
Best Bearer (DL) From “Coverage by Best Bearer (DL)”
Modulation (DL) From “Coverage by Best Bearer (DL)”
Peak RLC / Effective RLC / Application From “Coverage by Throughput (DL)”
Channel Throughput (kbps)
Peak RLC / Effective RLC / Application Cell From “Coverage by Throughput (DL)”
Capacity(kbps)

Uplink
PUSCH & PUCCH C/(I+N) Level (UL) (dB) From “Coverage by C/(I+N) Level (UL)”
Transmission power (UL) (dBm) From “Coverage by C/(I+N) Level (UL)”
Best Bearer (DL) From “Coverage by Best Bearer (UL)”
Modulation (DL) From “Coverage by Best Bearer (UL)”
Peak RLC / Effective RLC / Application From “Coverage by Throughput (UL)”
Channel Throughput (kbps)
Peak RLC / Effective RLC / Application Cell From “Coverage by Throughput (UL)”
Capacity(kbps)

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 138
Monte-Carlo Simulation Studies
 LTE Simulation Process – Overview
Start

Mobile Generation & Distribution

 The process is executed in five
Best Server Determination (RSRP) major steps
User Rejection

RSRP <
Min. RSRP?
Y No Coverage  Mobile Generation & Distribution :
N
Atoll generates user distribution of
Downlink Calculation
(SS, PBCH, PDSCH, PDCCH)
each simulation based on traffic
distribution weighted by Poisson
PDSCH
SINR < Min.?
Y No DL Service
distribution
N
Uplink Calculation
(PUSCH, PUCCH)  Determination of Best Server : Atoll
determines best server for each
SINR < Min.? Y No UL Service
mobile based on downlink RSRP
N
N
Bearer Determination & Channel
Throughput Calculation  Downlink Calculation : This
Scheduling & Radio Resource includes SS, PBCH, PDSCH &
Management
PDCCH calculation
# Users < Scheduler
Y
Max. User? Saturation

N
 Uplink Calculation : This includes
Min.
Resource
PUCCH & PUSCH calculation
Thp Demand N
Saturation
Met?
Y  RRM & Cell Load Calculation : This
User Throughput Calculation
includes determination of RB
Converged?
requirement, their allocation,
throughput determination and load
Y

Generate Results
estimation

Stop

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 140
 LTE Simulation Parameters

 Recommended Settings
 Number of simulations : 20
 Max. DL traffic load : 100%
 Max. UL traffic load : 100%
 Global scaling factor : 1 (to use the same traffic map)
 Check the box for the specific traffic map to be used for analysis

 Sensitivity of the parameters : High

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 141
 Simulation Convergence

 In Atoll, MC simulation convergence may be achieved by any of the

following
 Max. Number of Iterations: The maximum number of iterations that Atoll
should run to make convergence

 DL Traffic Load Convergence Threshold: The relative difference in terms of

downlink traffic load that must be reached between two iterations

 UL Traffic Load Convergence Threshold: The relative difference in terms of

uplink traffic load that must be reached between two iterations

 UL Noise Rise Convergence Threshold: The relative difference in terms of

uplink noise rise that must be reached between two iterations

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 142
 Convergence Definition in Atoll
Simulation Convergence Criteria

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 143
 Simulation Results (1/5)

 Atoll provides both summarized and detailed results:

 Key Summary Statistics

 Total number and percentage of users unable to connect: rejected users, and
the number of rejected users per rejection cause.
 The number and percentage of users connected to a cell, the number of users
per activity status, and the total UL and DL throughputs they generate.

Simulation Statistics

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 144
 Simulation Results (2/5)

 Key Site Level Statistics

 Peak / Effective RLC Aggregate Throughput for individual services (UL / DL)

 Effective Application Throughput (UL / DL)

 User Rejection per cause type (number & percentages)

 Connection type (DL-only, UL-only, UL+DL) with statistics

Simulation Statistics

 User plots can be generated

based on throughput / failure
etc
 Simulation results can also
be committed back to the
cells
 Based on user level
simulation results, new Atoll
prediction plots can also be
generated

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 145
 Simulation Results (3/5)

 Key Cell Level Statistics

 Peak/Effective RLC Aggregate Thpt; Effective Application Thpt (UL / DL)

 User Rejection per cause type (number & percentages)

 Connection type (DL-only, UL-only, UL+DL) with statistics

 Cell Loading (UL / DL) and UL noise rise (not available at site level)
Simulation Statistics

 User plots can be generated

based on throughput / failure
etc
 Simulation results can also
be committed back to the
cells
 Based on user level
simulation results, new Atoll
prediction plots can also be
generated

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 146
 Simulation Results (4/5)

 Key Mobile Statistics

 Location, mobility, activity, indoor/outdoor, serving cell, path loss

 RSRP, RSRQ, SS, PBCH, PDSCH, PUSCH, BLER statistics

 Peak / Effective RLC / Channel / User Throughput (UL / DL)

 Allocated bandwidth, bearer index

Simulation Statistics

 User plots can be generated

based on throughput / failure
etc
 Simulation results can also
be committed back to the
cells
 Based on user level
simulation results, new Atoll
prediction plots can also be
generated

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 147
 Simulation Results (5/5)

 Atoll rejects users for the following reasons:

 No coverage
 In this case, the RSRP detected by a user is below the “Min RSRP” level
set for the best serving cell
 No Downlink Service
 In this case, the user PDSCH SINR experienced by a user is less than the
minimum value defined for the lowest bearer
 No Uplink Service
 Similarly, the uplink SINR experienced by a user is less than the minimum
value defined for the lowest bearer
 Scheduler / Resource Saturation
 If the user is not selected by the scheduler for resource allocation (max.
user limit reached), it will be termed as scheduler saturation
 If all of the cell’s resources were used up by other users such that min.
throughput demand could not be met or if, for an uplink user, the minimum
uplink throughput demand was higher than the uplink allocated bandwidth
throughput, it will be termed as resource saturation

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 148
Neighbor List Creation
 Neighbor List Creation

 Atoll is capable of performing automatic neighbor list allocation.

 Sensitivity : High. Without neighbor list, automatic PCI allocation does not work

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 150
 Configuring Importance

 Atoll can automatically

assign importance to the
neighbors.
 Sensitivity : Medium.
Neighbor list importance will
help in RF optimization and
performance improvement

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 151
Physical Cell ID Creation
 Automatic PCI Creation

 Atoll is capable of performing automatic PCI allocation. Following settings are recommended.
 SSS ID : same per site
 Existing neighbor relation : checked
 Min. reuse distance ; checked (5 miles minimum)
 Neighbor list : should be generated in advance

 Sensitivity : High. Without PCI, ICIC analyses will not work

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 153
Summary – Parameter List
 Transmitter Table

Parameter Recommended Setting

Transmitter Type Intra-network (Server and Interferer)
Number of Transmission Antenna Ports 1 (for no MIMO); 2 (for 2x2 MIMO); 4 (for 4x4 MIMO)
Number of Reception Antenna Ports 2
Noise Figure (dB) 3 (should follow vendor's spec)
BTS Equipment eNode-B (as defined in the BTS equipment table)
TMA Equipment TMA (as defined in the TMA table)
Feeder Equipment Feeder (as defined in the feeder table)
Transmission Feeder Length (ft) As per feeder length of the sector
Reception Feeder Length (ft) As per feeder length of the sector
Miscellaneous Transmission Losses (dB) 0.5
Miscellaneous Reception Losses (dB) 0.5
Main Propagation Model As per tuned propagation model
Main Calculation Radius (miles) 18
Main Resolution (m) 25 (or as per digital map resolution)
Default Cyclic Prefix 0-Normal
PDCCH Overhead (symbols) 4 (BW<=3MHz); 3 (BW=5,10 MHz); 2 (BW=> 15 MHz)
PUCCH Overhead RBs: 5MHz = 2, > 5MHz = 4.
Reference Signal EPRE 0-Calculated from Max. Power & EPRE Offset
Uplink Power Adjustment Margin (dB) 1
Adaptive MIMO Switching Criterion Reference Signal C/N

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 155
 Cell Table
Parameter Recommended Setting
Active TRUE
Layer (Lowest Layer = Highest Priority) 0
Frequency Band Freq. Band of operation
Channel Number channel number of operation
Physical Cell ID PCI ID
Max Power (dBm) Max. PA power
RS EPRE (dBm) Calculated
SS EPRE Offset / RS (dB) 0
PDSCH EPRE Offset / RS (dB) 0
PBCH EPRE Offset / RS (dB) 0
PDCCH EPRE Offset / RS (dB) 0
Min RSRP (dBm) -121 (700MHz band); should follow 3GPP spec
AMS & MU-MIMO Threshold (dB) -19
ICIC Delta Path Loss Threshold (dB) 5
Fractional Power Control Factor 1
Max UL Noise Rise (dB) 6
Max PUSCH C/(I+N) (dB) 21
Interference Coordination Support None
LTE Equipment Cell UL Rx Equipment (as per vendor specification)
Scheduler Proportional Fair (or as per vendor specification)
Diversity Support (DL) SU-MIMO
Diversity Support (UL) Receive Diversity
Traffic Load (DL) (%) 100 (or as per prediction requirement)
ICIC Ratio (DL) (%) 0
Traffic Load (UL) (%) 100 (or as per prediction requirement)
UL Noise Rise (dB) 6
ICIC UL Noise Rise (dB) 0
MU-MIMO Capacity Gain (UL) 1
Inter-technology DL Noise Rise (dB) 0
Inter-technology UL Noise Rise (dB) 0
Max Traffic Load (DL) (%) 100
Max Traffic Load (UL) (%) 100
Max Number of Users 2000
Max number of intra-technology neighbours 16
Max number of inter-technology neighbours 16

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 156
 Scheduler Table

Proportional
Name Max C/I Proportional Fair Round Robin
Demand
1 - Proportional
Scheduling Method 3 - Max C/I 0 - Proportional Fair 2 - Round Robin
Demand

Target Throughput for 2 - Application 2 - Application 2 - Application 2 - Application

Voice Services Throughput Throughput Throughput Throughput

Target Throughput for 0 - Peak RLC 0 - Peak RLC 0 - Peak RLC 0 - Peak RLC
Data Services Throughput Throughput Throughput Throughput

Bearer Selection
0 - Bearer Index 0 - Bearer Index 0 - Bearer Index 0 - Bearer Index
Criterion

Uplink Bandwidth
2 - Best Bearer 2 - Best Bearer 2 - Best Bearer 2 - Best Bearer
Allocation Target

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 157
 MUG Curve

Number of Users MUG Comment

1 1
2 1
3 1
4 1
5 1
6 1
7 1
8 1
9 1
10 1
11 1
12 1
13 1
14 1 MUG is used with
proportional fair
15 1
scheduler only. Unless
16 1
specified by the vendors,
17 1 MUG should be set to 1
18 1
19 1
20 1
21 1
22 1
23 1
24 1
25 1
26 1
27 1
28 1
29 1
30 1

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 158
 Terminal Parameter Setting

Parameter UE – MIMO UE – MIMO UE – Non-

Handset Data Card MIMO

Min Power -40 -40 -40

Max Power 23 21 23

Gain (dBi) 0 0 0

Losses (dB) 5 5 5

Noise Figure (dB) 8 8 8

DL Rx DL Rx DL Rx
Reception Equipment
Equipment Equipment Equipment
Antenna

Diversity Support MIMO MIMO (none)

No. of MIMO Tx Ant 1 1 1

No. of MIMO Rx Ant 2 2 1

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 160
 Neighbor Creation & Configuration

Parameter Recommended Setting

Min. inter-site distance (miles) 10

Max. number of neighbors 16

Percentage of minimum covered area 5%

Resolution (m) 25

Minimum RSCP (dBm) -121 (700MHz band); shall follow 3GPP specs

RSRP margin (dB) 5

Shadowing taken into account Unchecked

Indoor coverage Unchecked

Coverage factor (Min. / Max. importance) 1% / 30%

Adjacency factor (Min. / Max. importance) 30% / 60%

Co-site factor (Min. / Max. importance) 60% / 100%

Take into account co-site factor Checked

Take into account adjacency factor Checked

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 161
 Automatic PCI Creation

Parameter Recommended Setting

Neighbor List Should be present in advance

SSS ID Same per site

Existing neighbor relationship Checked

Minimum reuse distance (miles) 5

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 162
 Monte-Carlo Simulation

Parameter Recommended Setting

Number of Simulations 20

Max DL Traffic Load 100%

Max UL Traffic Load 100%

Global Scaling Factor 1

Max. Number of Iterations 100

DL Traffic Load Convergence Threshold 1%

UL Traffic Load Convergence Threshold 1%

UL Noise Rise Convergence Threshold 0.5 dB

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 163
 Clutter Properties
Model standard C/I Standard Indoor Loss SU-MIMO Gain Additional Diversity Additional Diversity
Code Name Height (ft)
deviation (dB) Deviation (dB) (dB) Factor Gain (DL) (dB) Gain (UL) (dB)
1 Core Urban 35.28 7.7 1 20 1 0 0
2 Dense Urban 14.29 7.7 1 20 1 0 0
3 Urban 14.89 7.7 1 20 1 0 0
4 Commercial_Industrial 10.14 7.7 1 20 1 0 0
5 Suburban few Trees 21.7 6.7 1 15 1 0 0
6 Suburban with Trees 22.76 6.7 1 15 1 0 0
7 Residential few Trees 15.5 6.7 1 15 1 0 0
8 Residential with Trees 16.55 6.7 1 15 1 0 0
9 Rural few Trees 5.39 6.9 1 6 1 0 0
10 Rural with Trees 7.43 6.7 1 6 1 0 0
11 Convention Center 19.05 7.7 1 20 1 0 0
12 Major Stadium 15.21 6.9 1 15 1 0 0
Minor Stadium_Theme
13 14.76 6.9 1 6 1 0 0
Park_Fairgrounds
14 High School Building 15.14 6.9 1 15 1 0 0
15 University_College Building 17.91 7.7 1 20 1 0 0
16 Airport Runway 3.28 6.9 1 6 1 0 0
17 Airport Terminal 13.12 7.7 1 20 1 0 0
18 Airport Building 16.4 6.7 1 15 1 0 0
19 Airport Rural 3.28 6.7 1 6 1 0 0
20 Primary Road 0.55 6.7 1 6 1 0 0
21 Secondary Road 2.5 6.7 1 6 1 0 0
22 Tertiary Road 7.7 6.7 1 6 1 0 0
23 Other Paved_Impervious 9.45 6.7 1 6 1 0 0
Golf Course_Park_Urban
24 3.28 6.9 1 6 1 0 0
Recreation
25 Open In Urban 3.28 6.9 1 6 1 0 0
26 Open Suburban Residential 3.28 6.9 1 6 1 0 0
27 Open Rural 3.28 6.9 1 6 1 0 0
28 Grassland_Rangeland 0.28 6.9 1 6 1 0 0
29 Cultivated Cropland 1.28 6.9 1 6 1 0 0
30 Scrub Vegetation 2.9 6.9 1 6 1 0 0
31 Shrubland_Woodland 0.91 6.9 1 6 1 0 0
32 Deciduous Forest 20.31 6.9 1 6 1 0 0
33 Coniferous Forest 29.59 6.9 1 6 1 0 0
34 Mixed Forest 45.54 6.9 1 6 1 0 0
35 Forested Wetland 44.21 6.9 1 6 1 0 0
36 Wetland 19.38 6.9 1 6 1 0 0
37 Inland Water 0 6.9 1 6 1 0 0
38 Sea Water 0 6.9 1 6 1 0 0

80-W2838-4 Rev A Qualcomm Proprietary & Confidential MAY CONTAIN U.S. AND INTERNATIONAL EXPORT CONTROLLED INFORMATION 164
Questions?