Professional Documents
Culture Documents
Issue 01
Date 2014-04-26
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: http://www.huawei.com
Email: support@huawei.com
Contents
2 Overview.........................................................................................................................................3
2.1 Introduction....................................................................................................................................................................3
2.2 Benefits...........................................................................................................................................................................3
2.3 Application Scenarios.....................................................................................................................................................3
3.6.2 Principle.....................................................................................................................................................................13
3.6.3 Monitoring.................................................................................................................................................................13
6 Parameters.....................................................................................................................................21
7 Counters........................................................................................................................................25
8 Glossary.........................................................................................................................................26
9 Reference Documents.................................................................................................................27
1.1 Scope
This document describes eNodeB flow control, including its technical principles, related
features, network impact, and engineering guidelines.
Any managed objects (MOs), parameters, alarms, or counters described herein correspond to
the software release delivered with this document. Any future updates will be described in the
product documentation delivered with future software releases.
This document applies only to LTE FDD. Any "LTE" in this document refers to LTE FDD, and
"eNodeB" refers to LTE FDD eNodeB.
eRAN7.0 01 (2014-04-26)
This issue does not include any changes.
2 Overview
2.1 Introduction
With flow control, an eNodeB controls input and output flow to prevent overload and improve
equipment stability. Flow control includes control-plane, user-plane, OAM data flow control,
and uplink synchronized UE number control.
l An eNodeB controls input flow to prevent overload of the eNodeB and ensure the eNodeB's
processing capability when services increase dramatically.
l An eNodeB controls output flow to prevent overload of the peer network element (NE).
2.2 Benefits
When services increase dramatically, flow control decreases the possibility of eNodeB reset and
deteriorating access and handover success rates, thereby improving eNodeB stability.
According to data flow and load, there are eight load points in LTE networks, as shown in Table
2-1.
User-plane data Uplink service data flows Load point 5: An eNodeB is overloaded if UEs
flow from UEs to an eNodeB send too much service data to the eNodeB.
OAM data flow Uplink OAM data flows Load point 7: The OSS is overloaded if an
from an eNodeB to the eNodeB sends too much OAM data to the OSS.
operations support
system (OSS)
This document describes flow control at the preceding eight load points.
This chapter describes the objectives, principles, and monitoring methods of control-plane flow
control.
3.1.1 Objective
The objective of MME overload-triggered flow control is to relieve the impact of MME overload
caused by a large number of UEs requesting access to the network. MME overload-triggered
flow control corresponds to load point 1 in Figure 2-1.
3.1.2 Principle
When an MME is overloaded, the MME sends an OVERLOAD START message to eNodeBs,
instructing the eNodeBs to start flow control. The eNodeBs then limit the number of UEs
accessing the network. When the MME is no longer overloaded, the MME sends an
OVERLOAD STOP message to the eNodeBs, instructing them to stop flow control. For details
about MME overload-triggered flow control, see 3GPP TS 36.413 Release 9 or Release 10.
Figure 3-1 shows the signaling exchange between an MME and an eNodeB in MME overload-
triggered flow control.
After receiving an OVERLOAD START message, the eNodeB processes access requests
according to 3GPP TS 36.413 Release 9 or Release 10 as follows:
If the Overload Action IE in the Overload Response IE within the OVERLOAD START
message is set to
-"reject RRC connection establishments for non-emergency mobile originated data
transfer" (i.e., reject traffic corresponding to RRC cause "mo-data" and
"delayTolerantAccess" in TS 36.331 [16]), or
-"reject RRC connection establishments for signalling" (i.e., reject traffic
corresponding to RRC cause "mo-data", "mo-signalling" and "delayTolerantAccess" in
TS 36.331 [16]), or
-"only permit RRC connection establishments for emergency sessions and mobile
terminated services" (i.e., only permit traffic corresponding to RRC cause
"emergency" and "mt-Access" in TS 36.331 [16]), or
-"only permit RRC connection establishments for high priority sessions and mobile
terminated services" (i.e., only permit traffic corresponding to RRC cause
"highPriorityAccess" and "mt-Access" in TS 36.331 [16]), or
-"reject only RRC connection establishment for delay tolerant access" (i.e., only
reject traffic corresponding to RRC cause "delayTolerantAccess" in TS 36.331 [16]).
The eNodeB shall ensure that only signalling traffic corresponding to permitted RRC
connections is sent to the MME.
3.1.3 Monitoring
Table 3-1 lists the performance counters related to MME overload-triggered flow control. For
details, see eNodeB Performance Counter Reference.
3.2.1 Objective
The objective of initial access request control is to relieve the impact of eNodeB overload caused
by a large number of UEs' access requests. A large number of UEs' access requests lead to heavy
load and even causes the eNodeB to reset. Initial access request control corresponds to load
points 5 and 6 in Figure 2-1.
3.2.2 Principle
Initial access requests include RRC connection requests, S1 handover requests, and X2 handover
requests. An initial access request is the start of an access procedure. After an eNodeB accepts
an initial access request, a lot of subsequent processing is required, causing numerous overheads.
Therefore, initial access request control can reduce the eNodeB load from the beginning.
To ensure high-priority services, initial access request control processes access requests based
on priorities. Access requests with the following causes are prioritized in descending order:
1. Emergency
2. High priority
3. Handover
4. CS paging
5. PS paging
6. Mobile-terminated access
7. Mobile-originated signaling
8. Mobile-originated data
9. Delay Tolerant Access
When an eNodeB is overloaded, the eNodeB rejects or discards some initial access requests
based on the CPU usage of the main control board as follows:
l When the CPU usage of the main control board is equal to or greater than 80% and less
than 90% (not configurable, with 5s as the smooth filtering value), and the CPU usage is
increasing, the eNodeB starts initial access request control based on the priorities of initial
access requests. The eNodeB sends an RRC Connection Reject message to a UE if the
eNodeB rejects the access request.
l When the CPU usage of the main control board is equal to or greater than 90%, the eNodeB
discards initial access request messages to prevent a breakdown and the eNodeB does not
send RRC Connection Reject messages to UEs.
l When the CPU usage of the main control board drops to less than 80%, the eNodeB stops
initial access request control based on the priorities of initial access requests.
l When the CPU usage of the main control board is equal to or greater than 80% and less
than 90%, and the CPU usage is decreasing, the eNodeB sends an RRC Connection Reject,
S1 HANDOVER FAILURE, or X2 HANDOVER FAILURE message to a UE if the
eNodeB rejects the access request. The RRC Connection Reject message contains the RRC
Reject Wait time IE. For details, see 3.4 RRC Reject Wait time.
If the eNodeB load is heavy for a long time, the eNodeB controls the number of signaling
messages received from peer NEs to reduce the load as follows:
l The eNodeB decreases the SCTP buffer threshold to reduce the number of signaling
messages sent from the MME to the eNodeB, thereby reducing the load in the downlink.
For details, see SCTP Congestion Control Feature Parameter Description.
l The eNodeB reduces the frequency of UEs' access to the network using Access Class (AC)
Barring to reduce the load in the uplink. For details about AC Control, see 3.5 AC
Control.
3.2.3 Monitoring
For details, see 3.1 MME Overload–triggered Flow Control.
3.3.1 Objective
The objective of paging message flow control is to relieve the impact of eNodeB overload caused
by a large number of paging messages. A large number of paging messages lead to heavy load
and even causes the eNodeB to reset. Paging message flow control corresponds to load point 3
in Figure 2-1.
3.3.2 Principle
A paging message is the start of a paging procedure. A large number of successfully processed
paging messages lead to a large number of UEs' access to the network and many system
overheads. Therefore, paging message flow control reduces the eNodeB load from the beginning.
To ensure high-priority services, paging message flow control processes paging messages with
different causes differentially. 3GPP Release 8, Release 9, and Release 10 define paging causes
as follows:
l According to 3GPP Release 10, the Paging Priority IE is used to distinguish between CS
and PS services. According to 3GPP Release 10, an eNodeB determines the paging
priorities of only CS fallback (CSFB) and IP multimedia subsystem enhanced multimedia
priority service (IMS-eMPS), which have high paging priorities to ensure user experience
for CS services. Paging priorities of PS services need to be planned by operators and
configured on the core network. The following table describes the Paging Priority IE.
When an eNodeB is overloaded, the eNodeB discards paging messages based on the CPU usage
of the main control board as follows:
l When the CPU usage of the main control board is equal to or greater than 80% (not
configurable, with 5s as the smooth filtering value), the eNodeB starts flow control based
on service priorities described in 3.2 Initial Access Request Control. The eNodeB
preferentially discards PS paging messages in paging message flow control.
l When the CPU usage of the main control board is less than 80%, the eNodeB stops flow
control based on the service priorities described in 3.2 Initial Access Request Control.
3.3.3 Monitoring
Table 3-2 lists the performance counters related to paging message flow control. For details,
see eNodeB Performance Counter Reference.
3.4.1 Objective
The objective of sending the RRC Reject Wait time IE is to prevent signaling bursts caused by
UEs' frequent access to the network. When an eNodeB rejects a UE's access request, the eNodeB
includes the RRC Reject Wait time IE in the RRC Connection Reject message to inform the UE
of the wait time before the UE initiates another access request. The function of sending the RRC
Reject Wait time IE corresponds to load point 5 in Figure 2-1.
3.4.2 Principle
When an eNodeB is overloaded, the eNodeB includes the RRC Reject Wait time IE in the RRC
Connection Reject message to inform the UE of the wait time before the UE initiates another
access request. The UE processes the RRC Reject Wait time IE according to 3GPP TS 36.331.
Figure 3-2 shows the signaling exchange between the eNodeB and UE.
When the CPU usage of the main control board is equal to or greater than 80% (not configurable,
with 5s as the smooth filtering value), the eNodeB starts initial access request control and
includes the RRC Reject Wait time IE in the RRC Connection Reject message.
The value of the RRC Reject Wait time IE is configurable. You can run the MOD
RRCCONNSTATETIMER command to set the RrcConnStateTimer.T302 parameter. For
details, see eNodeB MML Command Reference.
3.4.3 Monitoring
For details, see 3.1 MME Overload–triggered Flow Control.
3.5 AC Control
3.5.1 Objective
As defined in 3GPP TS 36.331, access class (AC) control is a method used to control the UE
access to the network. The objective is to relieve the impact of eNodeB overload caused by a
large number of UEs requesting access to the network. AC control corresponds to load point 5
in Figure 2-1.
3.5.2 Principle
With AC control, the eNodeB broadcasts AC control parameters to UEs in a cell through System
Information Block 2 (SIB2) messages. UEs then determine whether they can access the cell
based on the received AC control parameters. Figure 3-3 shows the signaling exchange between
the eNodeB and UE for AC control.
Figure 3-3 Signaling exchange between the eNodeB and UE for AC control
l Static AC control
With static AC control, after AC control parameters are configured on the OSS by an
operator, the eNodeB broadcasts parameters to UEs through system information (SI)
update, without considering the current network condition. The eNodeB implements static
AC control on the following objects:
– Emergency
– Mobile-originated data
– Mobile-originated signaling
– Multimedia telephony voice
You can adjust the frequencies of UEs' access to the network by running the MOD
CELLACBAR command on the eNodeB side and configuring the parameters
CellAcBar.AcBarringFactorForCall, CellAcBar.AcBarringFactorForSig,
CellAcBar.AcBarFactorForMVoice, CellAcBar.AcBarFactorForMVideo, and
CellAcBar.AcBarFactorForCsfb.
l Intelligent AC control
With intelligent AC control, the eNodeB automatically triggers or cancels AC control by
adjusting the mobile-originated signaling and mobile-originated data access frequencies
based on the load of a cell. For details, see Access Class Control Feature Parameter
Description.
3.5.3 Monitoring
For details, see 3.1 MME Overload–triggered Flow Control.
3.6.1 Objective
The objective of uplink synchronized UE number control is to relieve the impact of eNodeB
overload caused by too many uplink synchronized UEs in a short time in special scenarios such
as festivals. Uplink synchronized UE number control corresponds to load point 5 in Figure
2-1.
3.6.2 Principle
When there are too many UEs in a cell, the eNodeB selects some uplink synchronized UEs that
have not transmitted or received data for a period and changes the status of these UEs to out of
synchronization. This releases physical uplink control channel (PUCCH) and sounding reference
signal (SRS) resources for UEs that attempt to access or are handed over to the cell.
You can run the MOD ENODEBFLOWCTRLPARA command to specify the thresholds of
uplink synchronized UE number and duration without transmitting or receiving data by setting
the eNodeBFlowCtrlPara.AdaptUnsyncUserNumThd and
eNodeBFlowCtrlPara.AdaptUnsyncTimerLen parameters, respectively. For details, see
eNodeB MML Command Reference.
3.6.3 Monitoring
Table 3-3 lists the performance counters related to uplink synchronized UE number control. For
details, see eNodeB Performance Counter Reference.
This chapter describes the objectives, principles, and monitoring methods of user-plane flow
control.
4.1.1 Objective
The objective of downlink user-plane flow control is to ensure eNodeB reliability when the
amount of downlink user-plane data exceeds the eNodeB's processing capability. Downlink user-
plane flow control corresponds to load points 4 and 6 in Figure 2-1.
4.1.2 Principle
Downlink user-plane flow control reduces the number of downlink packets, including packets
carried by guaranteed bit rate (GBR) and non-GBR bearers. An eNodeB preferentially performs
flow control on packets carried by non-GBR bearers. If congestion persists after the eNodeB
has performed flow control on all packets carried by non-GBR bearers, the eNodeB performs
flow control on packets carried by GBR bearers until the eNodeB is no longer congested.
4.1.3 Monitoring
Table 4-1 lists the performance counters related to downlink user-plane flow control. For details,
see eNodeB Performance Counter Reference.
This chapter describes the objectives, principles, and monitoring methods of OAM flow control.
OAM includes software upgrades, file downloads and uploads, logs, MML commands, alarms,
performance counters, and performance monitoring data. OAM requires a large amount of data
transmission and CPU usage, which negatively impacts the control-plane and user-plane data
transmission during busy hours. The priorities of most OAM tasks are lower than those of
control-plane and user-plane data transmission. OAM flow control releases CPU resources for
control-plane and user-plane data transmission.
5.1.1 Objective
The objective of OAM flow control is to prevent OAM tasks from occupying too many resources
and negatively affecting control-plane and user-plane data transmission. OAM flow control
corresponds to load points 7 and 8 in Figure 2-1.
5.1.2 Principle
When services are busy, an eNodeB performs flow control to tasks of log recording, software
management, and performance monitoring based on service priorities to ensure that the key
performance indicators (KPIs) are maintained.
The priorities of control-plane, user-plane, and OAM services are described as follows in
descending order:
1. High-priority OAM tasks, including tasks related to MML commands, alarms, performance
counters, and high-priority logs
2. Control-plane and user-plane data transmission
3. Medium-priority OAM tasks, including tasks related to medium-priority logs
4. Low-priority OAM tasks, including tasks related to low-priority logs, performance
monitoring, and software download
NOTE
When an eNodeB is overloaded, the eNodeB performs OAM flow control as follows:
l The eNodeB does not perform flow control to high-priority OAM tasks.
l When the CPU usage of the main control board is equal to or greater than 70% (not
configurable, with 5s as the smooth filtering value), the eNodeB rejects new performance
monitoring tasks. The eNodeB also stops ongoing performance monitoring tasks.
l The eNodeB stops log recording based on log priorities. When the CPU usage of the main
control board is equal to or greater than 70%, the eNodeB stops low-priority log recording.
When the CPU usage of the main control board is equal to or greater than 80%, the eNodeB
stops medium-priority log recording.
l When the CPU usage of the main control board is less than 70%, the eNodeB accepts new
performance monitoring tasks.
l When the CPU usage of the main control board is less than 80%, the eNodeB restarts
medium-priority log recording. Similarly, when the CPU usage of the main control board
is less than 70%, the eNodeB restarts low-priority log recording.
5.1.3 Monitoring
OAM flow control interrupts performance monitoring tasks and a message indicating that tasks
are stopped because of flow control is displayed on the U2000.
6 Parameters
RrcCon T302 MOD LBFD-0 RRC Meaning: Indicates the length of timer T302. T302
nStateTi RRCCO 02007 / Connect specifies the time during which a UE whose RRC
mer NNSTA TDLBF ion connection request is rejected has to wait before the UE
TETIM D-00200 Manage can initiate a request again. This timer is started when
ER 7 ment the UE receives an RRCConnectionReject message and
LST stopped when the UE enters the RRC_CONNECTED
RRCCO mode or performs cell reselection.
NNSTA GUI Value Range: 1~16
TETIM Unit: s
ER
Actual Value Range: 1~16
Default Value: 4
CellAcB AcBarri MOD LBFD-0 Broadca Meaning: Indicates the access probability factor for
ar ngFactor CELLA 02009 / st of mobile-originated calls. A mobile-originated call is
ForCall CBAR TDLBF system granted access if the random number generated by the
LST D-00200 informat UE is less than this access probability factor; otherwise,
CELLA 9 ion the access request is rejected. According to 3GPP TS
CBAR 36.331, if any of the parameters AC11BarforCall,
AC12BarforCall, AC13BarforCall, AC14BarforCall,
and AC15BarforCall is set to BOOLEAN_TRUE, the
eNodeB sends UEs P00 as the access probability factor
for mobile-originated calls in the system information
block type 2 (SIB2), regardless of the actual setting of
the AcBarringFactorForCall parameter.
GUI Value Range: P00(0%), P05(5%), P10(10%), P15
(15%), P20(20%), P25(25%), P30(30%), P40(40%),
P50(50%), P60(60%), P70(70%), P75(75%), P80
(80%), P85(85%), P90(90%), P95(95%)
Unit: %
Actual Value Range: P00, P05, P10, P15, P20, P25, P30,
P40, P50, P60, P70, P75, P80, P85, P90, P95
Default Value: P95(95%)
CellAcB AcBarri MOD LBFD-0 Broadca Meaning: Indicates the access probability factor for
ar ngFactor CELLA 02009 / st of signaling. Signaling from a UE is granted access if the
ForSig CBAR TDLBF system random number generated by the UE is less than this
LST D-00200 informat access probability factor; otherwise, the access request
CELLA 9 ion is rejected. According to 3GPP TS 36.331, if any of the
CBAR parameters AC11BarForSig, AC12BarForSig,
AC13BarForSig, AC14BarForSig, and
AC15BarForSig is set to BOOLEAN_TRUE, the
eNodeB sends UEs P00 as the access probability factor
for signaling in the system information block type 2
(SIB2), regardless of the actual setting of the
AcBarringFactorForSig parameter.
GUI Value Range: P00(0%), P05(5%), P10(10%), P15
(15%), P20(20%), P25(25%), P30(30%), P40(40%),
P50(50%), P60(60%), P70(70%), P75(75%), P80
(80%), P85(85%), P90(90%), P95(95%)
Unit: %
Actual Value Range: P00, P05, P10, P15, P20, P25, P30,
P40, P50, P60, P70, P75, P80, P85, P90, P95
Default Value: P95(95%)
CellAcB AcBarF MOD LBFD-0 Broadca Meaning: Indicates the access probability factor for
ar actorFor CELLA 02009 / st of multimedia telephony (MMTEL) voice services. An
MVoice CBAR TDLBF system MMTEL voice service is granted access if the random
LST D-00200 informat number generated by the UE is less than this access
CELLA 9 ion probability factor; otherwise, the access request is
CBAR barred.
GUI Value Range: P00(0%), P05(5%), P10(10%), P15
(15%), P20(20%), P25(25%), P30(30%), P40(40%),
P50(50%), P60(60%), P70(70%), P75(75%), P80
(80%), P85(85%), P90(90%), P95(95%)
Unit: %
Actual Value Range: P00, P05, P10, P15, P20, P25, P30,
P40, P50, P60, P70, P75, P80, P85, P90, P95
Default Value: P95(95%)
CellAcB AcBarF MOD LBFD-0 Broadca Meaning: Indicates the access probability factor for
ar actorFor CELLA 02009 / st of multimedia telephony (MMTEL) video services. An
MVideo CBAR TDLBF system MMTEL video service is granted access if the random
LST D-00200 informat number generated by the UE is less than this access
CELLA 9 ion probability factor; otherwise, the access request is
CBAR barred.
GUI Value Range: P00(0%), P05(5%), P10(10%), P15
(15%), P20(20%), P25(25%), P30(30%), P40(40%),
P50(50%), P60(60%), P70(70%), P75(75%), P80
(80%), P85(85%), P90(90%), P95(95%)
Unit: %
Actual Value Range: P00, P05, P10, P15, P20, P25, P30,
P40, P50, P60, P70, P75, P80, P85, P90, P95
Default Value: P95(95%)
CellAcB AcBarF MOD LBFD-0 Broadca Meaning: Indicates the access probability factor for CS
ar actorFor CELLA 02009 / st of fallback (CSFB) services. If the random number
Csfb CBAR TDLBF system generated by a UE is less than the parameter value, the
LST D-00200 informat eNodeB accepts the CSFB service access request;
CELLA 9 ion otherwise, the eNodeB rejects the access request.
CBAR GUI Value Range: P00(0%), P05(5%), P10(10%), P15
(15%), P20(20%), P25(25%), P30(30%), P40(40%),
P50(50%), P60(60%), P70(70%), P75(75%), P80
(80%), P85(85%), P90(90%), P95(95%)
Unit: %
Actual Value Range: P00, P05, P10, P15, P20, P25, P30,
P40, P50, P60, P70, P75, P80, P85, P90, P95
Default Value: P95(95%)
eNodeB AdaptU MOD None None Meaning: Indicates the threshold of the ratio of uplink-
FlowCtr nsyncUs ENODE synchronized UEs in a cell or a baseband processing
lPara erNumT BFLOW unit. If the ratio of uplink-synchronized UEs in a cell or
hd CTRLP a baseband processing unit exceeds this threshold, the
ARA eNodeB adaptively enables some UEs that do not
LST transmit or receive data for a period of time greater than
ENODE the AdaptUnsyncTimerLen parameter value to enter the
BFLOW asynchronization state. As a result, PUCCH resources
CTRLP can be released and used by other UEs.
ARA GUI Value Range: 50~100
Unit: %
Actual Value Range: 50~100
Default Value: 100
7 Counters
8 Glossary
9 Reference Documents