Professional Documents
Culture Documents
Multi-Carrier Unified Scheduling (eRAN19.1 - 07)
Multi-Carrier Unified Scheduling (eRAN19.1 - 07)
Issue 07
Date 2024-01-05
and other Huawei trademarks are trademarks of Huawei Technologies Co., Ltd.
All other trademarks and trade names mentioned in this document are the property of their respective
holders.
Notice
The purchased products, services and features are stipulated by the contract made between Huawei and
the customer. All or part of the products, services and features described in this document may not be
within the purchase scope or the usage scope. Unless otherwise specified in the contract, all statements,
information, and recommendations in this document are provided "AS IS" without warranties, guarantees
or representations of any kind, either express or implied.
The information in this document is subject to change without notice. Every effort has been made in the
preparation of this document to ensure accuracy of the contents, but all statements, information, and
recommendations in this document do not constitute a warranty of any kind, express or implied.
Website: https://www.huawei.com
Email: support@huawei.com
Contents
1 Change History.........................................................................................................................1
1.1 eRAN19.1 07 (2024-01-05)..................................................................................................................................................1
1.2 eRAN19.1 06 (2023-11-30)..................................................................................................................................................1
1.3 eRAN19.1 05 (2023-09-27)..................................................................................................................................................2
1.4 eRAN19.1 04 (2023-08-23)..................................................................................................................................................3
1.5 eRAN19.1 03 (2023-06-29)..................................................................................................................................................3
1.6 eRAN19.1 02 (2023-04-28)..................................................................................................................................................4
1.7 eRAN19.1 01 (2023-03-09)..................................................................................................................................................5
1.8 eRAN19.1 Draft A (2022-12-30)........................................................................................................................................ 6
3 Overview................................................................................................................................. 14
4 Multi-carrier Unified Scheduling....................................................................................... 16
4.1 Principles.................................................................................................................................................................................. 16
4.1.1 Virtual Grid Function........................................................................................................................................................ 16
4.1.1.1 Virtual Grid Model Building........................................................................................................................................17
4.1.1.2 Virtual Grid Model Updating..................................................................................................................................... 24
4.1.1.3 Switches for the Virtual Grid Function................................................................................................................... 25
4.1.1.4 Conditions for Using Virtual Grid Models..............................................................................................................27
4.1.1.5 Relationship Between Virtual Grid Models and Handovers............................................................................ 29
4.1.2 Smart Carrier Selection................................................................................................................................................... 30
4.1.2.1 Smart Carrier Selection Scenarios and Procedures............................................................................................ 33
4.1.2.2 Carrier or Carrier Combination Priority Factors and Selection.......................................................................44
4.1.3 Ultra-Low-Latency Scheduling for Intra-eNodeB CA............................................................................................ 51
4.1.3.1 Delayed SCell Deactivation......................................................................................................................................... 51
4.1.3.2 Data Split with Ultra-Low Latency...........................................................................................................................52
4.1.4 Ultra-Low-Latency Scheduling for Inter-eNodeB CA Based on Relaxed Backhaul..................................... 53
4.1.5 Multi-Frequency Optimal Carrier Scheduling.......................................................................................................... 54
4.2 Network Analysis.................................................................................................................................................................. 55
4.2.1 Benefits................................................................................................................................................................................. 55
4.2.2 Impacts.................................................................................................................................................................................. 57
4.2.2.1 Smart Carrier Selection................................................................................................................................................ 57
4.2.2.2 Ultra-Low-Latency Scheduling for Intra-eNodeB CA......................................................................................... 70
4.2.2.3 Ultra-Low-Latency Scheduling for Inter-eNodeB CA Based on Relaxed Backhaul..................................73
4.2.2.4 Multi-Frequency Optimal Carrier Scheduling...................................................................................................... 74
4.3 Requirements......................................................................................................................................................................... 77
4.3.1 Licenses................................................................................................................................................................................. 77
4.3.2 Software................................................................................................................................................................................78
4.3.2.1 Smart Carrier Selection................................................................................................................................................ 78
4.3.2.2 Ultra-Low-Latency Scheduling for Intra-eNodeB CA......................................................................................... 79
4.3.2.3 Ultra-Low-Latency Scheduling for Inter-eNodeB CA Based on Relaxed Backhaul..................................79
4.3.2.4 Multi-Frequency Optimal Carrier Scheduling...................................................................................................... 80
4.3.3 Hardware.............................................................................................................................................................................. 83
4.3.4 Networking.......................................................................................................................................................................... 83
4.3.5 Others.................................................................................................................................................................................... 84
4.4 Operation and Maintenance............................................................................................................................................. 84
4.4.1 Data Configuration........................................................................................................................................................... 84
4.4.1.1 Data Preparation............................................................................................................................................................ 84
4.4.1.2 Using MML Commands............................................................................................................................................... 92
4.4.1.3 Using the MAE-Deployment...................................................................................................................................... 95
4.4.2 Activation Verification......................................................................................................................................................95
4.4.3 Network Monitoring......................................................................................................................................................... 96
5 Parameters.............................................................................................................................. 97
6 Counters.................................................................................................................................. 98
7 Glossary................................................................................................................................... 99
8 Reference Documents........................................................................................................ 100
1 Change History
Technical Changes
None
Editorial Changes
Added descriptions of the impact relationships between smart carrier selection
based on virtual grids and energy saving functions. For details, see 4.2.2.1 Smart
Carrier Selection.
Added descriptions of the impact relationships between ultra-low-latency
scheduling for intra-eNodeB CA and the following functions: energy saving based
on proactive scheduling and precise Scheduling for CA UEs. For details, see 4.2.2.2
Ultra-Low-Latency Scheduling for Intra-eNodeB CA.
Revised descriptions of the impact relationships between multi-frequency optimal
carrier scheduling and the downlink scheduling policy. For details, see 4.2.2.4
Multi-Frequency Optimal Carrier Scheduling.
Added descriptions of the network impacts of smart carrier selection based on
virtual grids. For details, see 4.2.2.1 Smart Carrier Selection.
Technical Changes
None
Editorial Changes
Corrected the criteria for determining the traffic volumes of a UE in event-
triggered carrier selection (triggered by traffic model changes). For details, see
Event-triggered Carrier Selection (Triggered by Traffic Model Changes).
Revised the descriptions of the impact relationship between smart carrier selection
based on virtual grids and frequency-priority-based inter-frequency handover. For
details, see 4.2.2.1 Smart Carrier Selection.
Technical Changes
Change Description Parameter Change RAT Base Station
Model
Added support for Activated the BIT2 option FDD ● 3900 and 5900
prohibition of smart of the TDD series base
carrier selection eNBCellRsvdPara.RsvdSw stations (macro
triggered by inter- Para10 parameter. base stations)
frequency event A2 ● DBS3900
for non-CA UEs under LampSite and
weak coverage. For DBS5900
details, see Event- LampSite
triggered Carrier
Selection (Triggered
by Event A1 or A2)
in 4.1.2.1 Smart
Carrier Selection
Scenarios and
Procedures.
Editorial Changes
Added setting notes and MML command examples for the
TRAFFIC_AWARENESS_SCS_OPT_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw parameter. For details, see
4.4.1.1 Data Preparation and 4.4.1.2 Using MML Commands.
Added the description that the eNodeB delivers event A4 measurement
configurations to a non-CA UE if the strongest inter-frequency neighboring cell
predicted using virtual grid models is an unknown neighboring cell. For details, see
4.1.1.4 Conditions for Using Virtual Grid Models.
Added descriptions of the network impacts of smart carrier selection based on
virtual grids. For details, see 4.2.2.1 Smart Carrier Selection.
Added descriptions of prohibition of SCell configuration triggered by inter-
frequency event A2 for CA UEs under weak coverage. For details, see Event-
triggered Carrier Selection (Triggered by Event A1 or A2) in 4.1.2.1 Smart
Carrier Selection Scenarios and Procedures.
Technical Changes
None
Editorial Changes
Revised descriptions in this document.
Revised the descriptions of the network impact of "ultra-low-latency scheduling
for intra-eNodeB CA". For details, see 4.2.2.2 Ultra-Low-Latency Scheduling for
Intra-eNodeB CA.
Revised the setting notes for the AI_ENHANCEMENT_OPT_SWITCH option of the
eNodeBResModeAlgo.ServiceMode parameter. For details, see 4.4.1.1 Data
Preparation.
Technical Changes
Change Description Parameter Change RAT Base Station
Model
Added prohibition of Activated the BIT13 option FDD ● 3900 and 5900
smart carrier selection of the TDD series base
for UEs that have eNBRsvdPara.RsvdSwPara stations (macro
accessed local cells 7 parameter. base stations)
through RRC ● DBS3900
connection LampSite and
reestablishments. For DBS5900
details, see 4.1.1.1 LampSite
Virtual Grid Model
Building and 4.1.2
Smart Carrier
Selection.
Added support for the Modified parameter: FDD ● 3900 and 5900
multi-carrier unified Added the series base
scheduling for CA UBBPD_SCH_FUNCTION_ stations (macro
function by the ENH_SWITCH option to base stations)
UBBPd. For details, the ● DBS3900
see 4.1.5 Multi- EnodebAlgoExtSwitch.Ne LampSite and
Frequency Optimal tworkPrfmOptSwitch DBS5900
Carrier Scheduling parameter. LampSite
and 4.3.3 Hardware.
Editorial Changes
Revised descriptions in this document.
Added descriptions of the benefits of the multi-frequency optimal carrier
scheduling function. For details, see 4.2.1 Benefits.
Added descriptions of the impact relationship between smart carrier selection
based on virtual grids and low power consumption mode. For details, see 4.2.2.1
Smart Carrier Selection.
Technical Changes
Change Description Parameter Change RAT Base Station
Model
Added support for cell Activated reserved FDD ● 3900 and 5900
capability adjustment parameters: TDD series base
during smart carrier ● eNBRsvdPara.RsvdPara stations (macro
selection when PCI- 62 base stations)
modulo-3 results of ● DBS3900
involved cells are the ● eNBCellRsvdPara.Rsvd
U8Para67 LampSite and
same. For details, see DBS5900
4.1.2.2 Carrier or LampSite
Carrier Combination
Priority Factors and
Selection.
Editorial Changes
Revised descriptions in this document.
Added the feature ID and feature name of the TDD feature TDLNOFD-19150601
Multi-Frequency Optimal Carrier Scheduling. For details, see 2.3 Features in This
Documentand 2.4 Feature Differences Between FDD and TDD.
Technical Changes
None
Editorial Changes
Revised descriptions in this document.
Technical Changes
Change Description Parameter Change RAT Base Station
Model
Editorial Changes
Revised descriptions in this document.
Revised descriptions of virtual grid model indexes. For details, see 4.1.1.1 Virtual
Grid Model Building.
Revised descriptions of the handling related to UE handover. For details, see
Special Handling of UE Handovers in 4.1.2.1 Smart Carrier Selection Scenarios
and Procedures.
Revised descriptions of the application scenarios of short-cycle carrier selection.
For details, see Periodic Carrier Selection in 4.1.2.1 Smart Carrier Selection
Scenarios and Procedures.
This document only provides guidance for feature activation. Feature deployment and
feature gains depend on the specifics of the network scenario where the feature is
deployed. To achieve optimal gains, contact Huawei professional service engineers.
Functions mentioned in this document work properly only when enabled in the
specified applicable scenarios (such as RAT and networking). If a function not
mentioned in this document is enabled or a function is enabled in a scenario not
specified as applicable, exceptions or other impacts may occur.
Software Interfaces
Any parameters, alarms, counters, or managed objects (MOs) described in this
document apply only to the corresponding software release. For future software
releases, refer to the corresponding updated product documentation.
For example, in the statement "TDD cells are compatible with enhanced MU-
MIMO", "TDD cells" indicates that this function cannot be used in non-TDD cells.
3 Overview
select optimal carriers or carrier combinations for UEs based on the UE-level
spectral efficiency obtained from the models. For details, see 4.1.1 Virtual
Grid Function.
● Smart carrier selection: During initial access of UEs, the base station considers
the bandwidths, load, and virtual-grid-predicted spectral efficiency of different
frequencies to select carriers or carrier combinations that can provide the
maximum downlink throughput for the UEs. Periodic or event-triggered
evaluation is performed for UEs that stay online for a long time, so that the
UEs are always served by the optimal carriers or carrier combinations. For
details, see 4.1.2 Smart Carrier Selection.
● Ultra-low-latency scheduling for intra-eNodeB CA: Delayed SCell deactivation
is supported. It prevents frequent SCell deactivation and activation and
therefore reduces the performance loss caused by SCell activation latency. In
addition, the number of interactions between the PCell and SCells is reduced
to allow for simultaneous scheduling in the PCell and activated SCells. For
details, see 4.1.3 Ultra-Low-Latency Scheduling for Intra-eNodeB CA.
● Ultra-low-latency scheduling for inter-eNodeB CA based on relaxed backhaul:
In the data split process of inter-eNodeB CA based on relaxed backhaul, there
is a relatively long delay before the buffered data is scheduled. Ultra-low-
latency scheduling enables the eNodeB to start data distribution and physical
uplink control channel (PUCCH) resource allocation to the SCells of an
involved CA UE as early as in the transmission time interval (TTI) in which the
buffer for the UE is no longer empty, increasing the user-perceived throughput
of the UE. In addition, a proxy Radio Link Control (RLC) entity can be set up
at each eNodeB serving an SCell of a UE involved in inter-eNodeB CA based
on relaxed backhaul. Each proxy RLC entity distributes data to the MAC layer
of the corresponding inter-eNodeB SCell, preventing inaccurate scheduling
capability estimation caused by the delay. For details, see 4.1.4 Ultra-Low-
Latency Scheduling for Inter-eNodeB CA Based on Relaxed Backhaul.
● Multi-frequency optimal carrier scheduling: This function includes two sub-
functions, scheduling priority optimization for CA and multi-carrier unified
scheduling for CA. For details, see 4.1.5 Multi-Frequency Optimal Carrier
Scheduling.
– In scheduling priority optimization for CA, the base station calculates the
scheduling capability of each carrier for individual CA UEs based on the
spectral efficiency, bandwidth, and load. It maintains a high scheduling
priority of a CA UE on the carrier with the highest scheduling capability
and adjusts the scheduling priorities of the UE on the carriers that do not
have the highest scheduling capability based on the scheduling delay to
prevent the UE from preempting resources of non-CA UEs.
– In multi-carrier unified scheduling for CA, the base station schedules each
CA UE on its primary component carrier (PCC) and secondary component
carriers (SCCs) in a unified manner so that the PCC and SCCs can obtain
the buffered traffic volume in real time and the PCC can allocate PUCCH
resources to the SCCs in real time. This increases scheduling accuracy for
CA UEs on their PCCs and SCCs and improves downlink user experience.
4.1 Principles
Virtual Grids
A virtual grid is a signal grid of a cell. UEs in the same area of a cell normally have
the same radio signal characteristics, including the reference signal received power
(RSRP) and other information of the serving cell and intra-frequency neighboring
cells. Based on the measurements of radio signals, UEs with the same radio
characteristics are categorized as a single group. The base station considers these
UEs to be located in one signal grid, which is referred to as a virtual grid, as
illustrated in Figure 4-1. For example, if the measurement result for a frequency
measured by UE 1 is [(F1-Cell 1, F1-RSRP 1), (F1-Cell 2, F1-RSRP 2), (F1-Cell 3, F1-
RSRP 3), ...] and the measurement result for the same frequency measured by UE
2 is the same as that of UE 1, UE 1 and UE 2 are considered to be in the same
virtual grid (grid 1).
Virtual grid models include RSRP prediction models and spectral efficiency
prediction models. The spectral efficiency prediction models are categorized into
intra-frequency models and inter-frequency models. After obtaining a UE's virtual
grid information on a frequency, the eNodeB queries the RSRP prediction model
for a neighboring frequency to quickly predict the RSRP of the UE on this
neighboring frequency. In addition, the eNodeB queries the spectral efficiency
prediction models for target frequencies to quickly predict the spectral efficiency
of the UE on these frequencies.
1. The eNodeB determines the scope of cells and frequencies for which virtual
grid models are to be built. The scope is determined based on the attributes
of the cells and frequencies, and each RSRP prediction model or spectral
efficiency prediction model is identified by an index in the format of [Cell,
Frequency].
The following cells will not be selected for virtual grid model building:
– Cells in which UEs are stationary and therefore handovers are not
involved, such as eMTC-only cells and NB-IoT cells (eMTC and NB-IoT are
short for enhanced Machine Type Communication and Narrowband
Internet of Things, respectively.)
– Cells with multiple physical cell IDs, for example, a single frequency
network (SFN) cell or a cell with the Cell.MultiRruCellFlag parameter
set to BOOLEAN_TRUE
– High-speed-mobility cells, for example, cells with high speed mobility
enabled, cells with ultra high speed mobility enabled, cells with the
PATTERN_ADJ_SWITCH option of the
SectorSplitGroup.SectorSplitSwitch parameter selected, and cells on the
same frequency as those with the BEAM_NUM_ADJ_SWITCH option of
the SectorSplitGroup.SectorSplitSwitch parameter selected
– Cells that can only serve as SCells, where UE admission is not supported
and therefore samples cannot be collected, for example, cells with the
Cell.WorkMode parameter set to DL_ONLY or LAA and cells with the
ForbidCellSiSwitch option of the CellSiMap.SiSwitch parameter selected
– Cells with cell radius greater than 100 km enabled
The following neighboring frequencies will not be selected for virtual grid
model building:
– Frequencies (identified by the EutranInterNFreq.DlEarfcn parameter)
that belong to band 46 and serve licensed assisted access (LAA) cells
– Frequencies for which the VG_MODEL_FORBID_BUILD_FLAG option of
the EutranInterNFreq.AggregationAttribute parameter is selected
– Frequencies for which the FREQ_MEAS_FLAG option of the
EutranInterNFreq.AggregationAttribute parameter is deselected
– Frequencies (identified by the EutranInterNFreq.DlEarfcn parameter) on
each of which at least one cell of the eNodeB is working with the
BEAM_NUM_ADJ_SWITCH option of the
SectorSplitGroup.SectorSplitSwitch parameter selected
The following UEs will not be selected for virtual grid model building:
– UEs for which smart carrier selection is not allowed. When the eNodeB
builds and maintains virtual grid models, it does not collect data of these
UEs, which can be specified based on SPIDs, UE capabilities, or QCIs.
UM ● If the
PTe AI_ENHANCEMENT_
SWITCH option of
the
eNodeBResModeAl-
go.ServiceMode
parameter is selected,
the eNodeB supports
90 RSRP prediction
models and 114
spectral efficiency
prediction models.
● If the main control
board is working in
LTE&NR (LN) or
GSM&UMTS<E&NR
(GULN) mode and
the
AI_ENHANCEMENT_
OPT_SWITCH option
of the
eNodeBResModeAl-
go.ServiceMode
parameter is selected,
the eNodeB supports
18 RSRP prediction
models and 36
spectral efficiency
prediction models.
● If both options are
deselected, the
eNodeB supports 36
RSRP prediction
models but no
spectral efficiency
prediction models.
If the number of virtual grid models to be built exceeds the limit, the models
are prioritized based on the following factors, and then higher-priority models
are selected for building as long as the limit allows.
a. Intra-eNodeB or inter-eNodeB model: If a model is to be built for the
operating frequency of an intra-eNodeB inter-frequency cell, the model is
an intra-eNodeB model, whereas if it is for the operating frequency of an
inter-eNodeB inter-frequency cell, the model is an inter-eNodeB model. If
the INTER_ENB_VG_MDL_PRI_BUILD_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw parameter is
deselected, intra-eNodeB models take precedence over inter-eNodeB
models. If this option is selected, inter-eNodeB models take precedence
over intra-eNodeB models.
b. Local cell bandwidth: If a model is to be built for a larger-bandwidth cell,
the model has a higher priority.
c. Neighboring frequency priority: The camping priority is considered. If a
model is to be built for a neighboring frequency with a higher camping
priority, the model has a higher priority.
d. Neighboring frequency bandwidth: If a model is to be built for a larger-
bandwidth neighboring frequency, the model has a higher priority. This
factor applies only to intra-eNodeB models.
Table 4-2 and Table 4-3 provide an example to illustrate how the preceding
factors work when the INTER_ENB_VG_MDL_PRI_BUILD_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw parameter is deselected.
F3 20 MHz 7
F3 20 MHz 7
The eNodeB does not build virtual grid models within 14 days of an eNodeB software
upgrade. However, models built before the upgrade can still be used during this
period. After the 14 days, the eNodeB restarts the model building process.
3. When data collection is complete, the eNodeB starts the model training
process and calculates the accuracy of each model.
4. The eNodeB checks the model accuracy to evaluate whether the conditions
for launching models are met. If the accuracy of an RSRP prediction model
meets a certain requirement, the model is considered to have been built
successfully and can be put into use. Otherwise, the RSRP model is considered
to have failed to be built and the eNodeB goes back to 2. If the number of
consecutive building failures has reached the maximum, the eNodeB waits for
28 days before it tries to build the model again. If the relative error of a
spectral efficiency prediction model meets a certain requirement, the eNodeB
considers that the conditions are met and the model has been built
successfully, and it puts the model into use. Otherwise, the eNodeB considers
that the model has failed to be built.
As the use of virtual grid models increases CPU usage, the eNodeB performs flow
control as follows: It stops using virtual grid models to inquire about the RSRP and
spectral efficiency of neighboring frequencies when board CPU usage exceeds
60%. It resumes the inquiry when the board CPU usage drops below 45%.
After virtual grid models are built, the eNodeB monitors their performance and
updates them as follows:
1. After a model is put into use, the eNodeB starts the model monitoring and
updating process, in which it performs 2 and 4 in parallel.
2. The eNodeB monitors related KPIs in real time, based on which the eNodeB
determines the changes in the handover success rate while the model is being
used. It then proceeds to 3.
3. If KPIs such as the success rate of inter-frequency handovers based on the
model-based prediction results are lower than expected for two consecutive
hours, the eNodeB suspends the use of this model and puts it into use again
one day later. If the KPIs are still lower than expected, the eNodeB goes to 6
to build a new model, and uses the new model after it is successfully built. If
the KPIs are higher than expected after the model is put into use again, the
model will continue to be used.
4. The eNodeB monitors the model accuracy by performing a model accuracy
test every third day and then proceeds to 5.
5. If the model accuracy decreases, the eNodeB stops using the model, proceeds
to 6 to build a new model, and uses the new model after it is successfully
built.
6. The eNodeB builds a new model. It randomly selects a specific number of UEs
at a specific time in the cell for which the new model is to be built, collects
data from these UEs, and uses the collected data to build the new model.
Virtual grid model updating also involves the following mechanisms:
When the VG_MDL_PRED_NCELL_ENH_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw parameter is selected and virtual
grid model building is triggered, the eNodeB considers the following inter-
frequency neighboring cells reported by UEs during data collection as unknown
neighboring cells and includes them in model training; after the virtual grid
models are built, the eNodeB can predict the RSRP values of these unknown
neighboring cells:
● Neighboring cells with physical cell identifier (PCI) conflicts
● Neighboring cells not in the inter-frequency neighboring relation table (NRT)
● Neighboring cells whose coverage areas dynamically change
These neighboring cells have the COVERAGE_DYNAMIC_CHANGE_FLAG
option of the EutranExternalCell.AggregationAttribute parameter selected.
This option can be manually set. It can also be automatically set by the
eNodeB based on the X2 interface, if such a neighboring cell is served by
another eNodeB and the BEAM_NUM_ADJ_SWITCH option of the
SectorSplitGroup.SectorSplitSwitch parameter is selected for the sector
corresponding to the cell.
that uses virtual grid models is enabled and is disabled only when all features that
use virtual grid models are disabled.
This document mainly focuses on smart carrier selection. For details about Mid-
and Low-Band Coordination, see Low-Band Booster.
Building of When any of the options that control the building of RSRP
spectral prediction models is selected, spectral efficiency prediction
efficiency models can be built if either of the following options is
prediction selected:
models ● LTE_SPEC_EFF_VG_MDL_BUILD_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw
parameter
● SMART_CARRIER_SELECTION_ENH_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw
parameter
Use of spectral When any of the options that control the use of RSRP
efficiency prediction models is selected, spectral efficiency prediction
prediction models can be used if the
models SMART_CARRIER_SELECTION_ENH_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw parameter is
selected.
● Condition 2
– For a candidate frequency for SCell configuration for a CA UE, the
prediction result meets the entering or leaving condition for event A5.
– For a candidate frequency for the PCC anchoring of a CA UE:
▪ If MultiCarrUnifiedSch.VgModelPredHoMeasRsrpThld is set to a
value other than ACCURACY100_ENH, the prediction result meets
the entering or leaving condition for event A5.
▪ If MultiCarrUnifiedSch.VgModelPredHoMeasRsrpThld is set to
ACCURACY100_ENH, the prediction result meets the leaving
condition for event A5.
– For a candidate frequency for the handover of a non-CA UE:
▪ If MultiCarrUnifiedSch.VgModelPredHoMeasRsrpThld is set to a
value other than ACCURACY100_ENH, the prediction result meets
the entering or leaving condition for event A4.
▪ If MultiCarrUnifiedSch.VgModelPredHoMeasRsrpThld is set to
ACCURACY100_ENH, the prediction result meets the leaving
condition for event A4.
The virtual grid models of the serving cell may not provide the spectral efficiency
of a UE on some neighboring frequencies. In this case, the eNodeB can use the
inter-frequency event A4/A5 measurement results of the UE to obtain the RSRP of
a neighboring frequency and then input the RSRP into the spectral efficiency
model of the cell on the neighboring frequency to query the spectral efficiency of
this frequency.
The following is an example in which the UE is served by cell 1 on F1. Cell 1 has
the following three neighboring frequencies:
● Neighboring frequency F2, on which cell 2 operates
● Neighboring frequency F3, on which cell 3 operates
● Neighboring frequency F4, on which cell 4 operates
Table 4-6 lists the virtual grid models built by the eNodeB.
In the preceding example of virtual grid models, the spectral efficiency models of
cell 1 for F1 and F2 are available. In this case, the eNodeB can obtain the RSRP of
the UE on F1 from A3 measurements and uses the obtained RSRP to query the
spectral efficiency of F1 and F2. In the example, the spectral efficiency models of
cell 1 for F3 and F4 are unavailable. In this case, the eNodeB can instruct the UE in
cell 1 to measure F3 so as to obtain the RSRP of the UE on F3. Then, the eNodeB
uses the obtained RSRP to query the spectral efficiency model of cell 3 for F3 or
F4, thus allowing the spectral efficiency of F3 or F4 to be obtained.
case, the eNodeB periodically checks the model-based prediction results based
on event A3 measurement reports from the UE. As long as at least one
candidate neighboring frequency meets the entering condition for event A4,
the eNodeB delivers event A4 measurement configurations to the UE and
performs a handover based on the measurement results.
● Unnecessary handovers
– Service-based inter-frequency handover
When the MultiCarrUnifiedSch.VgMdlPredNecHoMeasRsrpThld
parameter is set to ACCURACY100_ENH and smart carrier selection
based on virtual grids takes effect, the eNodeB does not deliver
measurement configurations related to any candidate neighboring
frequency if all these frequencies are predicted to meet the leaving
condition for the measurement event related to service-based inter-
frequency handovers. In this case, the eNodeB periodically checks the
model-based prediction results based on event A3 measurement reports
from the UE. As long as at least one candidate neighboring frequency
meets the entering condition for the measurement event related to
service-based inter-frequency handovers, the eNodeB delivers
configurations of the measurement event related to service-based inter-
frequency handovers to the UE and performs a handover based on the
measurement results.
– PRB-usage-based connected mode load equalization
Models are used in the same way as in service-based inter-frequency
handovers.
– User-number-based connected mode load equalization
Models are used in the same way as in service-based inter-frequency
handovers.
For details about coverage-based inter-frequency handover and service-based
inter-frequency handover, see Mobility Management in Connected Mode.
For details about PRB-usage-based connected mode load equalization and user-
number-based connected mode load equalization, see Intra-RAT Mobility Load
Balancing.
To allow more carrier combinations for CA to be configured and increase the inter-
frequency overlapping coverage area as well as the proportion of occasions where
smart carrier selection based on virtual grids takes effect, it is recommended that
smart carrier selection be enabled together with SCC coverage threshold
adaptation. For details about SCC coverage threshold adaptation, see Low-Band
Booster.
If prohibition of smart carrier selection for specific UEs is enabled, the eNodeB
performs adaptive PCC anchoring and SCell configuration for the CA UEs involved.
For details, see Carrier Aggregation. Specific UEs include:
● UEs for which smart carrier selection is not allowed. These UEs can be
specified based on SPIDs, UE capabilities, or QCIs.
– These UEs can be specified based on SPIDs by selecting the
SCS_FORBID_SW option of the SpidCfg.AggregationAttribute
parameter. For details, see Flexible User Steering.
– These UEs can be specified based on UE capabilities by selecting the
SCS_FORBID_SW option of the UeCompat.BlacklistControlExtSwitch2
parameter. For details, see Terminal Awareness Differentiation.
– These UEs can be specified based on QCIs by selecting the
SCS_FORBID_SW option of the QciPara.AggregationAttribute
parameter. For a UE running services with multiple QCIs, smart carrier
selection is not allowed as long as the SCS_FORBID_SW option of the
QciPara.AggregationAttribute parameter is selected for any of the QCIs.
During smart carrier selection based on virtual grids, the eNodeB performs special
handling of handovers in special scenarios.
Figure 4-5 takes initial access as an example to illustrate the signaling procedure.
Figure 4-5 Smart carrier selection based on virtual grids during initial access
For non-CA UEs, based on the UE-reported event A3 measurement results, the
eNodeB selects only the frequencies whose RSRP and spectral efficiency can be
predicted by virtual grid models as the candidate frequencies. If no virtual grid
models are applicable for a non-CA UE and the
LOAD_BASED_CARR_SEL_OPT_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchEnSw parameter is selected, the
eNodeB selects a neighboring cell whose downlink PRB usage difference from the
serving cell is greater than or equal to 15% based on A4 measurement results for
carrier selection.
NOTE
▪ If both the network and UE support NSA networking and an SCG has
been added for the UE, timer 1 is specified by the
NsaDcMgmtConfig.NsaDcLteScellConfigIntvl parameter.
NOTE
When timer 1 expires and the procedure for smart carrier selection based on
virtual grids is triggered, timer 2 restarts.
● Short-cycle carrier selection takes effect only for non-CA UEs and is used for
prompt selection of optimal carriers for UEs upon network load changes.
The cycle is 100 ms. At the end of each cycle, if the percentage of the data
transmission duration in the cycle in the cell is greater than or equal to 60%,
the eNodeB initiates the procedure for smart carrier selection based on virtual
grids and randomly selects three frequencies as the candidate frequencies for
each UE that meets all the following conditions: (1) The downlink traffic
volume of the UE is greater than 10 kilobytes; (2) the percentage of the data
transmission time of the UE in the total data transmission time of all UEs in
the cell is less than or equal to 50%; (3) the modulation and coding scheme
(MCS) index used for the UE is less than or equal to 20.
As short-cycle carrier selection increases CPU usage, the eNodeB performs
flow control by stopping short-cycle carrier selection when the board CPU
usage exceeds 70% and resuming the selection when the board CPU usage
drops below 45%.
model of the UE according to the uplink and downlink traffic volumes. The traffic
models are listed in Table 4-8.
Table 4-9 lists the criteria for determining the uplink and downlink traffic volumes
of a UE.
Figure 4-8 Smart carrier selection based on virtual grids when event A1 or A6 is
triggered
The A1-triggered procedure allows the eNodeB to quickly select optimal carriers or
carrier combinations for the UEs during their movement from low-frequency cell
edges to cell centers in a multi-frequency concentric cell networking scenario. (A
low-frequency cell is a cell whose downlink center frequency is less than 1 GHz.)
● When both PCell and SCell changes are allowed and the
EnhancedPccAnchorSwitch option of the ENodeBAlgoSwitch.CaAlgoSwitch
parameter is selected, the eNodeB delivers A1 measurement configurations to
UEs during initial access, necessary incoming handovers, or incoming RRC
connection reestablishments. The RSRP threshold for event A1 is specified by
the CaMgtCfg.EnhancedPccAnchorA1ThdRsrp parameter. After receiving an
A1 measurement report from a UE, the eNodeB sends an RRC Connection
Reconfiguration message to the UE to stop A1 measurements and triggers the
procedure for smart carrier selection based on virtual grids. If the
LOAD_BASED_CARR_SEL_OPT_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchEnSw parameter is selected and
PCell change has not been triggered for a CA UE, the eNodeB performs smart
carrier selection at 10-second intervals for the UE.
● When only SCell changes are allowed and the
CaMgtCfg.EnhancedSccSelA1ThldRsrp parameter is set to a value other
than –40 dBm, the eNodeB delivers A1 measurement configurations to UEs
during initial access, necessary incoming handovers, or incoming RRC
connection reestablishments, with the RSRP threshold for event A1 being the
value of this parameter. After receiving an A1 measurement report from a UE,
the eNodeB sends an RRC Connection Reconfiguration message to the UE to
stop A1 measurements and triggers the procedure for smart carrier selection
based on virtual grids.
The A6-triggered procedure allows the eNodeB to quickly select optimal carriers or
carrier combinations for the UEs when the PCell does not provide contiguous
coverage and inter-eNodeB cells cannot be configured as SCells. If the
A6_BASED_SCS_SW option of the
MultiCarrUnifiedSch.MultiCarrierUnifiedSchSw parameter is selected, the
eNodeB performs smart carrier selection for UEs whose SCells cannot be changed
after they send A6 measurement reports. This procedure takes effect only for CA
UEs and allows both PCells and SCells to be changed for the UEs.
for these UEs. For details, see "Mobility and Load Management" in Carrier
Aggregation.
Due to compatibility issues, certain UEs may experience a low handover success
rate or failures of FDD-to-TDD handovers with SCell configuration. To prevent
these problems from occurring, operators can set the
UeCompat.BlacklistControlExtSwitch1 parameter to enable differentiated
handling for these UEs:
● If the SCS_BASED_HO_SWITCH option is selected, handovers triggered by
smart carrier selection are prohibited for these UEs.
● If the SCS_FDD_TO_TDD_HO_WITH_SCC_SW option is selected, FDD-to-TDD
handovers with SCell configuration triggered by smart carrier selection are
prohibited for these UEs.
For details about differentiated handling, see Terminal Awareness Differentiation.
If the signal levels of the target cell for a handover and its intra-frequency
neighboring cell are similar, the handover may fail.
● For a low-frequency cell from which handovers for smart carrier selection
based on virtual grids are triggered, operators can set the
MultiCarrUnifiedSch.GFactorThldForLowFreqHoTrg parameter to a value
other than 32767. With this setting, an inter-frequency neighboring cell can
be selected as the target cell for the handovers only when it meets the
conditions specified by the parameter, thereby reducing the handover failure
probability.
● For a non-low-frequency cell from which handovers for smart carrier selection
based on virtual grids are triggered, operators can set the
MultiCarrUnifiedSch.GFactorThldForHoTrgCell parameter to a value other
than 32767. With this setting, an inter-frequency neighboring cell can be
selected as the target cell for the handovers only when it meets the
conditions specified by the parameter, thereby reducing the handover failure
probability.
virtual grid models can be used for smart carrier selection only when
the predictions cover all frequencies.
▪ The smart carrier selection algorithm considers cell load only when
the CaMgtCfg.MinDlAvgToBeScheduledUeNum parameter is set to
a non-zero value.
▪ If the UE does not support pairing for massive MIMO, the eNodeB
filters out the combinations that include high-load inter-frequency
massive MIMO cells acting as candidate SCells. A high-load inter-
frequency massive MIMO cell is defined as follows:
○ If the CaMgtCfg.HLUeCntThldForScellConfig parameter is set
to 65535, massive MIMO cells in the inter-frequency massive
MIMO load steering state are defined as high-load inter-
frequency massive MIMO cells. For details about inter-frequency
massive MIMO load steering, see Intra-RAT Mobility Load
Balancing.
○ If the CaMgtCfg.HLUeCntThldForScellConfig parameter is set
to a value other than 65535, massive MIMO cells whose number
of UEs exceeds this value are defined as high-load massive
MIMO cells.
● In scheduling priority optimization for CA, the base station calculates the
scheduling capability of each carrier for individual CA UEs based on the
spectral efficiency, bandwidth, and load. It maintains a high scheduling
priority of a CA UE on the carrier with the highest scheduling capability and
adjusts the scheduling priorities of the UE on the carriers that do not have the
highest scheduling capability based on the scheduling delay to prevent the UE
from preempting resources of non-CA UEs.
● In multi-carrier unified scheduling for CA, the base station schedules each CA
UE on its PCC and SCCs in a unified manner so that the PCC and SCCs can
obtain the buffered traffic volume in real time and the PCC can allocate
PUCCH resources to the SCCs in real time. This increases scheduling accuracy
for CA UEs on their PCCs and SCCs and improves downlink user experience.
4.2.1 Benefits
Smart carrier selection based on virtual grids, ultra-low-latency scheduling for
intra-eNodeB CA, and multi-frequency optimal carrier scheduling must be enabled
together. If only one of them is enabled, the optimal effect cannot be achieved.
The benefits depend on related factors as follows:
● The larger the number of carriers on the network and the larger the
bandwidth of each carrier, the greater the benefits.
● The benefits increase when the network is neither extremely lightly loaded
nor extremely heavily loaded.
● The higher the proportion of last TTIs, the greater the benefits. The
proportion is equal to (L.Thrp.Time.DL – L.Thrp.Time.DL.RmvLastTTI)/
L.Thrp.Time.DL.
● The higher the proportion of CA transmission time, the greater the benefits.
– When the CaInstantlyJudgeSwitch option of the
CaMgtCfg.CellCaAlgoSwitch parameter is selected, the proportion of CA
transmission time is equal to L.Thrp.Time.DL.RmvLastTTI.CAUser/
L.Thrp.Time.DL.RmvLastTTI.
– When the CaInstantlyJudgeSwitch option of the
CaMgtCfg.CellCaAlgoSwitch parameter is deselected, the proportion of
CA transmission time is equal to L.Thrp.Time.DL.RmvLastTTI.CAUser/
L.Thrp.Time.DL.RmvLastTTI/[(L.Thrp.bits.DL.CAUser/L.Thrp.bits.DL)/
(L.Traffic.User.PCell.DL.Avg/L.Traffic.User.Avg)].
Smart carrier selection based on virtual grids enables the eNodeB to select
optimal carriers or carrier combinations for UEs from multiple carriers based on
factors such as the UE-level spectral efficiency and load. Therefore, the average
downlink UE throughput increases.
Smart carrier selection based on virtual grids also enables the eNodeB to be aware
of traffic models of UEs so that carriers or carrier combinations that provide the
highest uplink and downlink throughput are selected from multiple candidates for
the UEs. However, this function of traffic model-aware smart carrier selection
slightly decreases the average downlink UE throughput across the network. When
operators intend to increase the uplink throughput of an individual UE running
large-packet uplink services, it is recommended that this function be enabled in
one of the following scenarios:
● FDD+TDD networking
● Multi-band FDD networking
● Uplink 2CC aggregation is enabled, and the numbers of aggregated carriers in
the downlink and uplink cannot simultaneously reach their respective
maximums due to UE limitations.
Ultra-low-latency scheduling for intra-eNodeB CA increases the SCell usage and
therefore raises the average downlink data rate of UEs, especially in scenarios
where service bursts are frequent.
Multi-frequency optimal carrier scheduling increases the average downlink UE
throughput on the entire network. The benefits depend on the following factors:
● The larger the number of carriers on the network and the larger the
bandwidth of each carrier, the greater the benefits.
● The larger the number of frequencies on the network and the greater the
load difference between frequencies, the greater the benefits.
● The benefits are greater in the case of medium or heavy load than in other
load cases. It is recommended that multi-frequency optimal carrier scheduling
be enabled for base stations that each have an average downlink PRB usage
greater than 20%.
● The higher the proportion of CA UEs whose SCells are configured on the same
BBP as the PCell, the greater the benefits. It is recommended that multi-
carrier unified scheduling for CA, a sub-function of multi-frequency optimal
carrier scheduling, be enabled for base stations on each of which the
proportion is greater than 70%. The proportion is equal to
L.Traffic.User.SCell.SameBBPWithPCell.DL.Avg divided by
L.Traffic.User.SCell.DL.Avg.
● The higher the proportion of CA UEs, the greater the benefits. It is
recommended that multi-carrier unified scheduling for CA, a sub-function of
multi-frequency optimal carrier scheduling, be enabled for base stations on
each of which the proportion is greater than 60%. The proportion is equal to
L.Traffic.User.PCell.DL.Avg divided by L.Traffic.User.Avg.
When the Multi-carrier Unified Scheduling feature is enabled on a multi-frequency
FDD network, it is recommended that the following functions be also enabled to
further increase the average downlink UE throughput:
4.2.2 Impacts
Network Impacts
Smart carrier selection based on virtual grids has the following impacts on the
network:
● The UE distribution across frequency bands changes. Therefore, the values of
band-specific counters and indicators also change.
● The CPU usage of the main control board increases because of frequent
virtual grid model query and optimal carrier or carrier combination
evaluation.
● After this function is enabled, more resources of the main control board are
consumed. If the number of UEs is too large, congestion may occur earlier
than expected, causing the RRC connection setup success rate to decrease and
the service drop rate to increase.
● The number of handovers increases because the eNodeB needs to select
better carriers for UEs whenever there are changes such as load and
interference changes on the frequency bands.
● The number of signaling messages over the air interface and the signaling
radio bearer (SRB) traffic increase. The average MCS index may decrease
because SRBs use MCSs with small indexes. In addition, RRC connection
reconfiguration messages may fail to be sent, which increases the service drop
rate and the number of connection reestablishments.
● Due to gap-assisted inter-frequency measurements by UEs selected by the
eNodeB for virtual grid model building and updating, the uplink throughput
and downlink throughput of these UEs decrease and network interference
indicators vary slightly. As only a few UEs are selected, the impact on the
average downlink UE throughput and average uplink UE throughput is
negligible when there are more than 15 online UEs in a cell.
● For fast-moving UEs that cannot be identified, the selected carriers or carrier
combinations may not be optimal throughout the service duration. As a
result, the throughput of such UEs may decrease.
● After the eNodeB has built virtual grid models, enhanced processing of inter-
frequency neighboring cells unknown to the eNodeB increases the success
rate of inter-frequency handovers based on virtual grids if there are unknown
neighboring cells. In addition, due to these unknown neighboring cells, the
value of the L.VirtualGrid.Model.InterFreq.MeasFree.Times counter may
slightly increase or decrease.
● The overall network resource consumption increases slightly. As a result, fewer
resources are available for UEs performing full buffer services, and the
throughput of these UEs may decrease slightly.
● In light load scenarios with a large difference in the bandwidths of frequency
bands, the positive gains of smart carrier selection based on virtual grids may
be less than the negative impacts brought by measurements. As a result, the
UE throughput decreases in such scenarios.
● The number of handovers for PCC anchoring of NSA UEs when this function
takes effect is greater than that when intelligent selection of serving cell
combinations takes effect. (The latter function is controlled by the
CaSmartSelectionSwitch option of the ENodeBAlgoSwitch.CaAlgoSwitch
parameter.)
● If there are abnormal UEs on the network, the number of RRC connection
reestablishments increases. The increase can be prevented by selecting the
SCS_FORBID_SW option of the SpidCfg.AggregationAttribute or
UeCompat.BlacklistControlExtSwitch2 parameter.
● The eNodeB does not build virtual grid models within 14 days of an eNodeB
software upgrade. However, models built before the upgrade can still be used
during this period. After the 14 days, the eNodeB restarts the model building
process. Therefore, within 14 days of the upgrade, the numbers of gap-
assisted inter-frequency measurements, air interface signaling messages, and
RRC connection reconfiguration messages may decrease, and the service drop
rate may also decrease, further leading to changes in the values of user-plane
performance indicators, such as the number of UEs, user-perceived rate, and
traffic volume.
● On a network where there are both FDD cells and TDD cells, the SCells for
individual UEs may change from FDD cells to TDD cells or from TDD cells to
FDD cells. If the FDD and TDD cells use different DRX policies, the values of
the L.Cdrx.Enter.Num (which measures the number of times UEs enter DRX)
and L.Cdrx.Exit.Num (which measures the number of times UEs exit DRX)
counters may increase.
Function Impacts
● Functions in the category "RAN functions"
CA may not be
configured for the
UE on the LTE side.
reselection policy
specific to the
SPID of a CA UE
has been
configured, PCC
anchoring will not
be triggered for
the UE during
initial access.
Network Impacts
Ultra-low-latency scheduling for intra-eNodeB CA has the following impacts on
the network:
● The traffic distribution across frequency bands changes. Therefore, the values
of band-specific performance indicators also change.
● The traffic distribution between CA UEs and non-CA UEs changes. Therefore,
performance indicators specific to CA UEs and those specific to non-CA UEs
also change.
● The CPU usage of the BBP increases. If the average CPU usage of the BBP
exceeds 50% or the maximum CPU usage of the BBP exceeds 65%, it is
recommended that this function not be enabled.
● The number of UEs for which CA is activated increases. In the case of inter-
base-station CA, the amount of scheduling information exchanged between
the PCC and inter-base-station SCCs increases, causing the usage of inter-
base-station transmission resources to increase correspondingly. The values of
the following related counters increases: L.TRPIP.X2UTxBytes and
L.TRPIP.X2URxBytes, which respectively measure the numbers of bytes
transmitted and received on the user plane of the X2 interface;
VS.FEGE.TxBytes and VS.FEGE.RxBytes, which respectively measure the
numbers of bytes transmitted and received over the FE/GE ports;
VS.FEGE.TxMaxSpeed and VS.FEGE.RxMaxSpeed, which respectively measure
the maximum Tx and Rx rates over the FE/GE ports. When the proportion of
UEs in the inter-base-station CA state is high, the usage of transmission
resources increases significantly.
● The volume of data distributed to SCells increases, which causes the following
impacts:
– The overhead of uplink and downlink control channels slightly increases.
– In some scenarios, NACKs and DTXs in feedback from CA UEs cannot be
distinguished from each other. As a result, the RBLER and service drop
rate increase.
● After this function is enabled, the downlink rate on the LTE side increases. As
a result, in EN-DC scenarios, more data is distributed from the PDCP layer to
the LTE side, causing changes in the downlink rates on the NR side.
Function Impacts
RAT Function Function Reference Description
Name Switch
Network Impacts
Ultra-low-latency scheduling for inter-eNodeB CA based on relaxed backhaul has
the following impacts on the network:
● The traffic distributions on frequency bands change. As a result, the values of
band-specific performance indicators also change.
● The CPU usage of the BBP slightly increases. When the CPU usage of the BBP
is greater than 60%, this function is not recommended.
● This function is not recommended when the entire network is heavily loaded,
for example, when the downlink PRB usage in each of the PCell and SCells
exceeds 70%.
● When the load of SCells for inter-eNodeB CA based on relaxed backhaul is not
heavy, the volume of data distributed to the SCells increases, which has the
following impacts:
– The uplink and downlink control channel resource consumption increases
slightly. As a result, fewer resources are available for unidentified UEs
performing full buffer services, and the throughput of these UEs may
decrease slightly.
– In some scenarios, NACKs and DTXs in feedback from CA UEs cannot be
distinguished from each other. As a result, the RBLER and service drop
rate increase.
● When the load of SCells for inter-eNodeB CA based on relaxed backhaul is
heavy, the traffic volumes in these SCells decrease.
Function Impacts
None
Network Impacts
Multi-frequency optimal carrier scheduling has the following impacts on the
network:
● The traffic distribution across frequency bands changes. Therefore, the values
of band-specific performance indicators also change.
● The traffic distribution between CA UEs and non-CA UEs changes. Therefore,
performance indicators specific to CA UEs and those specific to non-CA UEs
also change.
● The volume of data distributed to SCells increases, which causes the following
impacts:
– The uplink and downlink control channel resource consumption increases
slightly. As a result, fewer resources are available for unidentified UEs
performing full buffer services, and the throughput of these UEs may
decrease slightly.
– In some scenarios, NACKs and DTXs in feedback from CA UEs cannot be
distinguished from each other. As a result, the RBLER and service drop
rate increase.
Function Impacts
RAT Function Function Switch Reference Description
Name
4.3 Requirements
4.3.1 Licenses
Multi-carrier Unified Scheduling requires the feature licenses listed in Table 4-10
and capacity licenses listed in Table 4-11.
NOTE
If both FDD and TDD cells exist on the base station, apply for licenses for both FDD and
TDD.
4.3.2 Software
Prerequisite Functions
RAT Function Name Function Switch Reference
Prerequisite Functions
RAT Function Name Function Switch Reference
Prerequisite Functions
RAT Function Name Function Switch Reference
Prerequisite Functions
RAT Function Function Switch Reference Description
Name
4.3.3 Hardware
Boards
● Smart carrier selection
– Main control boards: UMPTb series or later boards are required.
– BBPs: no requirements
● Ultra-low-latency scheduling for intra-eNodeB CA
– Main control boards: UMPTb series or later boards are required.
– BBPs: The LBBP is incompatible with ultra-low-latency scheduling for
intra-eNodeB CA. In a base station equipped with both the UBBP and
LBBP, some CCs may be served by the UBBP while others by the LBBP.
This decreases the performance gains brought by this function.
● Ultra-low-latency scheduling for inter-eNodeB CA based on relaxed backhaul
– Main control boards: UMPTb series or later boards are required.
– BBPs: Ultra-low-latency scheduling for inter-eNodeB CA based on relaxed
backhaul does not take effect on any base station equipped with both
the UBBP and LBBP.
● Multi-frequency optimal carrier scheduling
– Main control boards: UMPTb series or later boards are required.
– BBPs:
Multi-carrier unified scheduling for CA takes effect only for CA UEs whose
SCells are configured on the same BBP as the PCell.
In TDD, multi-carrier unified scheduling for CA does not take effect when
the UBBPd or LBBP is used.
In FDD, multi-carrier unified scheduling for CA does not take effect when
the LBBP is used.
However, when the UBBPe is used, multi-carrier unified scheduling for CA
does not take effect for CA UEs in the FDD+TDD CA state even if the
serving FDD and TDD cells are all configured on the UBBPe.
RF Modules
No requirements
4.3.4 Networking
The network has at least two carriers.
4.3.5 Others
Smart carrier selection based on virtual grid is not recommended in the following
scenarios:
● In high-speed mobility scenarios such as high-speed railway dedicated
networks, it is recommended that this function be disabled. This is to prevent
network performance deterioration because of inaccurate cell capability
evaluation due to fast UE mobility.
● During key event assurance, it is recommended that this function be disabled.
This is to prevent network performance deterioration due to the increase in
the CPU usage and SRB traffic.
● During long-term network reconstruction, it is recommended that this
function be disabled or both the
MultiCarrUnifiedSch.VgModelPredHoMeasRsrpThld and
MultiCarrUnifiedSch.VgMdlPredNecHoMeasRsrpThld parameters be set to
ACCURACY100_OFFSET0 to prevent network performance deterioration
caused by a decrease in the prediction accuracy of virtual grid models.
● On a network where cells are frequently activated and deactivated, it is
recommended that the VG_MODEL_FORBID_BUILD_FLAG option of the
EutranInterNFreq.AggregationAttribute parameter be selected to forbid
virtual grid model building.
To verify activation of smart carrier selection, use one of the following methods:
5 Parameters
You can find the EXCEL files of parameter reference and used reserved parameter list for
the software version used on the live network from the product documentation delivered
with that version.
Step 2 On the Parameter List sheet, filter the Feature ID column. Click Text Filters and
choose Contains. Enter the feature ID, for example, LOFD-001016 or
TDLOFD-001016.
Step 3 Click OK. All parameters related to the feature are displayed.
----End
Step 1 Open the EXCEL file of the used reserved parameter list.
Step 2 On the Used Reserved Parameter List sheet, use the MO, Parameter ID, and BIT
columns to locate the reserved parameter, which may be only a bit of a parameter.
View its information, including the meaning, values, impacts, and product version
in which it is activated for use.
----End
6 Counters
The following hyperlinked EXCEL files of performance counter reference match the
software version with which this document is released.
● Node Performance Counter Summary: contains device and transport counters.
● eNodeBFunction Performance Counter Summary: contains all counters related
to radio access functions, including air interface management, access control,
mobility control, and radio resource management.
NOTE
You can find the EXCEL files of performance counter reference for the software version used
on the live network from the product documentation delivered with that version.
----End
7 Glossary
8 Reference Documents
1. Carrier Aggregation
2. Mobility Management in Connected Mode
3. Licensed Assisted Access (TDD)
4. Intra-RAT Mobility Load Balancing
5. Multi-band Optimal Carrier Selection
6. Intra-RAT Mobility Load Balancing
7. SFN
8. High Speed Mobility
9. Extended Cell Range
10. Massive MIMO (FDD)
11. eMTC
12. Energy Conservation and Emission Reduction
13. Scheduling