Transcript KPI Establishment

Introduction to TD-LTE KPIs
Contents





Overview of Key Performance Indicators (KPIs)
KPI Establishment
KPI Report Procedure
KPI Classification
KPI Analysis
KPI Overview


Definition:
 The definitions of Key Performance Indicator (KPI) vary by
industries.
 KPIs are important criteria used to evaluate the operation of
wireless networks, for example, wireless call completion rate and
wireless call drop rate.
 A KPI is a value calculated through a series of PIs or counters.
KPIs reflect the network running status. Therefore, KPI values are
objective criteria and important basis for evaluating network running
status and quality.
KPI Related Concepts

PI
Although Performance Indicator (PI) as a type of Measure of Performance
is less critical than the KPI, yet it is still used as a reference data to
evaluate the network operation. PI values are obtained through
corresponding counters.
 Counter
 Each count corresponds to a counter, and there are thousands of counters
in a system.
 Northbound Counter
 The counter value reported according to the northbound Operation

Management Center (OMC) requirements is computed and reported by a
local counter. Northbound OMC uses these counters to compute KPI
values.
Relationship Among KPIs, PIs, Counters, and
Northbound Counters (1)


All KPIs, PIs, and northbound counters are computed through original
counters, and the corresponding formula are clearly defined in related
standards of ZTE.
The correctness of original counters and formulas is a prerequisite for
accurate KPI results.
Counter
calculate
calculate
North Counter
calculate
PI
KPI
Relationship Among KPIs, PIs, Counters, and
Northbound Counters (2)

Counters --- Vegetables


All KPIs and PIs need to use counters as raw materials for processing
and production.
Northbound Counters --- Semi-finished Dishes

Counters reported to northbound counters must be reprocessed by
northbound counters before they become KPIs.
Relationship Among KPIs, PIs, Counters, and
Northbound Counters (3)

PIs---Appetizers


KPIs---Main Courses

7
Although a PI is more often than not ignored, yet it has
the potential to become a KPI.
KPIs are indicators that attract the most attention, and
they are the main course that makes or breaks a meal.
Contents





Overview of KPIs
KPI Establishment
KPI Report Procedure
KPI Classification
KPI Analysis
KPI Establishment (1)


Take the call drop rate of TD-SCDMA video service as an
example to explain the establishment process of a KPI:
For this KPI, the computing formula extracted from the
northbound OMC is as follows:


(RAB.RelReqCSperTraffic.Conv.<5><5>+IU.NbrRABCSRelIuConn
perTraffic.Conv.<5><5>)/(RAB. SuccEstabCSNoQueuing.
Conv.<5><5>+RAB.SuccEstabCSQueuing.Conv.<5><5>)
This KPI is calculated by using four northbound counters
obtained through counters, as shown below:




RAB.RelReqCSperTraffic.Conv.<5><5>=C350131088
IU.NbrRABCSRelIuConnperTraffic.Conv.<5><5>=C350131279
RAB. SuccEstabCSNoQueuing. Conv.<5><5>=C350100042
RAB.SuccEstabCSQueuing.Conv.<5><5>=C350100137
KPI Establishment (2)



For this KPI, the computing formula extracted from the OMC is as follows:
 (C350131088 + C350131279) / (C350100042 + C350100137)
ZTE defines the preceding formula as follows:
 (Number of RABs that are released by RNC request for CS domain and
segmented based on service rates + Number of RABs that are released by
RNC request for CS domain and correspond to lu) / Number of successful
RAB establishments without queuing for CS domain + Number of
successful RAB establishments with queuing for CS domain)
“TD-SCDMA video service call drop rate” is calculated by using the four
counters. Then how are these counter values determined? How does the
system gather statistics of these values?
Isn’t it clear to explain it
in this way?
How to Determine Counter Values?

The following questions arise as the entire system has thousands of original
counters:
 Where do counter values come from?
 How to ensure the consistency between counters and actual calls?
 How to determine the correctness of the counters?
 How to distinguish so many types of counters?
Counters are provided according to the signaling statistics in call flow.
RRC connection request counts
RRC connection setup success counts
Contents





Overview of KPIs
KPI Establishment
KPI Report Procedure
KPI Classification
KPI Analysis
KPI Report Procedure 1



As there are many different types of counters, the foreground does not need to
submit all counter values while computing a KPI.
Use measurement task to control what data needs to be reported;
The reported data needs to feedback to the specific level of OMC from which a
measurement task is delivered.
Northbound OMC
Measurement task is an activity which
allows users to define which
performance data needs to be
collected. Then foreground equipment
system will collect performance data
according to measurement task, and
the OMC will store the collected
performance data.
OMC（MINOS）
OMM
eNodeB
KPI Report Procedure 2
Delivering measurement tasks is like growing vegetables on the
RNC
PI
Return
counter
values
KPI
Counte
r
Northboun
d counter
Compute PI
and KPI
Reported to
northbound
counter
KPI Report Procedure 3
Northbound OMC needs to use counter
(RAB.RelReqCSperTraffic.Conv.<5><5>) to deliver RAB
measurement cluster.
RAB.RelReqCSperTraffic.Conv.<5><5>=C350131088
OMC needs to use counter (C350131088) to deliver RAB release
request statistics.
Measurement Objects and Measurement Types



A measurement task needs to specify
what and how to measure, so it is extremely important to m
anage counters.
Measurement objects include:
RncFunction, IubLink, UtranCell, Carrier, UtranRelation,
GsmRelation, IucsLinkRnc, and IupsLinkRnc.
Measurement types:

A measurement type Indicate a collection of the same type of
counters. To well manage counters, some related counters are
categorized into one measurement type in the system.
Contents





Overview of KPIs
KPI Establishment
KPI Report Procedure
KPI Classification
KPI Analysis
KPI Specifications

There are many KPIs to measure the performance
of a network. Operators defines what KPIs each
manufacturer must provide, and standardizes
algorithms.
KPIs may constantly
change.
Classification of KPIs

The TD-LTE KPIs defined currently can be divided into
several categories, and each category consists of lots of
KPIs.
KPI Category
Related KPIs
Coverage
RSRP, RS-CINR, and coverage.
Call Establishment
RRC connection success rate, E-RAB establishment success
rate, wireless call completion rate, and E-RAB establishment
contestation rate.
Call Hold
RRC connection call drop rate and E-RAB call drop rate.
Mobility Management
Intra-eNodeB handover success rate, X2 interface handover
success rate, S1 interface handover success rate, and intersystem handover success rate (handover among GSM, WCDMA,
TD, and CDMA).
Quality
Uplink and downlink block error rate and uplink and downlink
MAC layer retransmission rate.
System Resource
traffic index and radio resource utilization indexes.
Contents





Overview of KPIs
KPI Establishment
KPI Report Procedure
KPI Classification
KPI Analysis






Coverage Indexes
Call Setup Indexes
Call Retention Indexes
Mobility Management Indexes
Quality Indexes
System Resource Indexes
RSRP

RSRP indicates an absolute value of received signal strength. To some extent,
it can reflect distance between UEs and eNodeBs. Therefore, this KPI value
can measure the cell coverage. RSRP is defined as the linear average over
the power contributions of the resource elements (RE) that carry cell-specific
reference signals (RS) within the considered measurement frequency
bandwidth.

Assume RSRP threshold is A, then RSRP coverage index (percentage) = The
number of test point with RSRP≥A/Total number of drive test points .

Before computing the RSRP, exclude abnormal test points whose RSRP value
is far beyond the normal value.

Within coverage areas, the TD-LTE network coverage must have a more than
95% probability of RSRP > -105dBm.

When a test antenna is put on car roof, there must be a more than 95%
probability of RSRP>－95dBm.
RS-CINR

RS-CINR indicates carrier-to-interference ratio on RS channel
measured by UE. It is one of the key indexes to indicate channel
quality.

Carrier to Interference plus Noise Ratio (CINR) is the ratio of carrier to
the noise and interference, which is measured by UE in drive test. In
the emulation tool CNP, RS-CINR=serving cell RSRP /(neighbor cell
RSRP+N); where, N is the thermal noise power.

Within coverage areas, the TD-LTE network coverage must have a
more than 95% probability of RS-CINR >0dB.
PDCCH SINR

PDCCH SINR indicates the quality of PDCCH.
 Formula: PDCCH SINR = (PDCCH receive power
of the best serving cell/ PDCCH interference in the
cell).
 The quality requirement of PDCCH can be met if
PDCCH SINR>-1.6dB.
Coverage

Wireless network coverage reflects the availability of a network.

RSRPR, and RSRQ  S, where:

RSRP indicates downlink reference signal received power;

RSRQ indicates the signal quality of received reference signal;

R and S indicate RSRP and RSRQ thresholds. Before calculation,
exclude abnormal tested points whose RSRP value or RSRQ value is
far beyond the normal value.

Within coverage areas, the TD-LTE network coverage must have a
more than 95% probability of RSRP  -105dBm, and RSRQ > -13.8dB.
Contents





KPI Overview
KPI Establishment
KPI Report Procedure
Index Classification
Key Index Analysis






Coverage Indexes
Call Setup Indexes
Call Retention Indexes
Mobility Management Indexes
Quality Indexes
System Resource Indexes
Call KPIs

Call success rate is one of the most important KPIs that reflect LTE system
performance and attract operators’ attention. A complete call completion rate
includes: Paging success rate, RRC connection setup rate, and E-RAB
assignment setup success rate.
UE
eNB
MME
SGW
PDN GW
RRC connection
setup
(1) RRC Connection Request
(2) RRC Connection Setup
The mobileoriginated process
does not include
paging;
After E-RAB
assignment
succeeds,
UE can use data
services such as
webpage browsing
and FTP
downloading.
(3) RRC Connection Complete
(4) Service Request (NAS signal)
(5) Initial Contest Setup Request / Attach Accept
E-RAB assignment
setup
(6) RRC Connection Reconfiguration
(7) RRC Connection Reconfiguration Complete
(8) Initial Contest Setup Response
(9) Update Bearer Request
(10) Update Bearer Request
(11) Update Bearer Response
(12) Update Bearer Response
RRC Connection Setup Success Rate

This index indicates the capacity of an eNodeBs or a cell to admit
UEs. RRC Connection setup success means that there is a signaling connection between a UE and
the network. After receiving a UE RRC Connection request, the eNodeB decides whether to set up an
RRC Connection for the UE.
Service-dependent
There are two
types of RRC
connection setup

Service-independent
The mobile-originated and mobileterminated calls
Emergency calls, and inter-system
cell update and registration
RRC connection setup success rate = Number of RRC connection setup successes/RRC connection
setup attempts × 100%.
RRC Connection
request attempts
RRC Connection
setup complete
1. There are corresponding counters in
place between two signaling points to
collect the number of RRC connection
request attempts.
2. When calculating service-dependent
RRC connection setup success rate,
collect statistics of only the number of
service-dependent RRC connection
setup successes.
E-RAB Connection Setup Success Rate




E-RAB connection is used to transfer voice, data and multmedia
services between UEs and CN. E-RAB connection setup success rate
reflects the capacity of an eNodeB or a cell to admit services.
E-RAB setup is initiated by the CN. A basic service is set up after ERAB connection is established.
E-RAB setup is implemented in the following two scenarios:
 When an eNodeB sets up the initial context for UE.
 When an MME initiates an E-RAB Setup Request (bearer setup)
E-RAB setup success rate = (Number of E-RAB setup successes (all
QCIs) /Number of E-RAB setup requests (all QCIs)) * 100%
E-RAB Setup in Initial Context Setup Procedure
UE
eNB
MME

The message that
contains the number of ERAB setup requests:
INITIAL CONTEXT
SETUP REQUEST

The message that
contains the number of ERAB setup successes:
INITIAL CONTEXT
SETUP RESPONSE
RRCConnectionRequest
RRCConnectionSetup
RRCConnectionSetupComplete(containing Attach
request)
E-RAB Connection Setup
Request
INITIAL UE MESSAGE (containing Attach request)
INITIAL CONTEXT SETUP REQUEST (containing
Attach Accept, Activate default EPS bearer context
request)
RRCConnectionReconfiguration (containing Attach
accept, Activate default EPS bearer context
request)
RRCConnectionReconfigurationComplete
E-RAB Connection Setup
Response
ULInformationTransfer(containing Attach
INITIAL CONTEXT SETUP RESPONSE
Complete,
Activate default EPS bearer context accept)
UPLINK NAS TRANSPORT(containing Attach
Complete,
Activate default EPS bearer context accept)
MME-Initiated E-RAB Setup
UE
Setup
Request
eNB

The message that contains the
number of E-RAB setup requests:
E-RAB SETUP REQUEST

The message that contains the
number of E-RAB setup successes:
E-RAB SETUP RESPONSE
MME
E-RAB
(1) E-RAB Setup Request /
Session Management Request
(2) RRC Connection Reconfiguration
(3) RRC Connection Reconfiguration Complete
(4) E-RAB Setup Response
E-RAB Setup Success
Radio Call Completion Rate

This index reflects the capability a cell has to
admit UE-originated calls, which directly affects
user network experience.
 Radio call completion rate = E-RAB setup success
rate  RRC connection setup success rate
(service-related)  100%
E-RAB Setup Blocking Rate

This index only reflects E-RAB setup failures caused by the system capacity,
namely, the E-RAB setup failure with cause value of radio resources not
available.
 E-RAB blocking rate ＝ Number of E-RABs that the MME refuses to admit
/Number of E-RABs for which the eNodeB requests an admission × 100%.
Signaling points used to collect statistical information are as follows:
(a) Failure to set up initial context
(b) E-RAB setup failure occurring when
radio resources are not available
UE
eNB
MME
Number of E-RABs admitted
Number of E-RABs admitted
(1) E-RAB Setup Request /
Session Management Request
(2) RRC Connection Reconfiguration
Number of E-RABs rejected
(3) RRC Connection Reconfiguration Complete
(4) E-RAB Setup Response
Number of E-RABs rejected
Contents





KPI Overview
KPI Establishment
KPI Report Procedure
Index Classification
Key Index Analysis






Coverage Indexes
Call Setup Indexes
Call Retention Indexes
Mobility Management Indexes
Quality Indexes
System Resource Indexes
Call Drop Rate for RRC Connection Exceptions

UE RRC connections may be released due to an eNodeB exception. This
probability indicates the RRC connection retention performance of an eNodeB
and reflects user network experience to some extent.
 Call drop rate for RRC connection exceptions = Number of RRC connection
releases due to an exception/(RRC connection success count + RRN
reconnection success count) × 100%
 The number of RRC connection releases due to an exception refers to the
number of all improper RRC connection releases. RRC connections can be
properly released when the core network properly releases RRC connections,
actively initiates the LOADBALANCE release, and initiates a RESET message.
UE
eNB
RRCConnectionRelease
The RRC Connection Release message carries
the cause. 3GPP TS 36.331 defines the cause
as the load balance or others.
Radio Interface Bearer Release – Normal Procedure 1
UE
eNB
MME
UE CONTEXT RELEASE COMMAND
RRCConnectionRelease
2
UE CONTEXT RELEASE COMPLETE




1
3
E-RAB Release(by MME)with UecontextRelease Command ：
Sampling Point 1: The eNodeB receives the UE CONTEXT RELEASE
COMMAND message from the MME.
Sampling Point 2: The eNodeB sends the RRC CONNECTION RELEASE
message to the UE.
Sampling Point 3: The eNodeB sends the UE CONTEXT RELEASE
COMPLETE message to the MME.
Radio Interface Bearer Release – Normal Procedure 2
Normal procedure: OMC signaling
Radio Interface Bearer Release – Normal Procedure 3
UE
eNB
MME
RESET
RESET ACKNOWLEDGE
RRCConnectionRelease




1
2
3
E-RAB Release(by MME)with S1 reset ：
Sampling Point 1: The eNodeB receives the RESET message from the MME.
Sampling Point 2: The eNodeB sends the RESET ACKNOWLEDGE message
to the MME.
Sampling Point 3: The eNodeB sends the RRC CONNECTION RELEASE
message to the UE.
Radio Interface Bearer Release – Normal Procedure 4
Normal procedure: OMC signaling
E-RAB Call Drop Rate

Generally, the eNodeB initiates a request to the MME for releasing one or
multiple E-RABs when any exception occurs.

In the case of UE losses, UE deactivations, or eNodeB exceptions, the
eNodeB sends a request to the MME for releasing the UE context. This may
cause the MME to release all setup E-RABs as well. The E-RAB call drop rate
relates to the number of eNodeB-initiated E-RAB release requests not due to
UE inactivation and the number of E-RABs released due to UE context release
initiated by eNodeB exceptions.

E-RAB call drop rate = (Number of service-based E-RABs (all QCIs) for which
the eNodeB sends a release request – Number of E-RABs (all QCIs) for which
the eNodeB sends a release request due to UE inactivation/Number of
successfully set up service-based E-RABs (all QCIs) + Number of cells that
hand over to the E-RAB (all QCIs) – Number of cells that hand over from the
E-RAB (all QCIs) × 100%
Contents





KPI Overview
KPI Establishment
KPI Report Procedure
Index Classification
Key Index Analysis






Coverage Indexes
Call Setup Indexes
Call Retention Indexes
Mobility Management Indexes
Quality Indexes
System Resource Indexes
TD-LTE Handover

LTE networks support manual handover only. UEs have to disconnect the links
with their serving cells and then get admitted to the target cells.
 LTE systems support intra-system and inter-system handovers. Intra-system
handovers can be classified into intra-frequency and inter-frequency
handovers according to carrier configuration conditions. Inter-system
handovers support the handover between all systems, for example 2G, 3G
(CDMA, WCDMA, and TD-SCDMA) .
 LTE handovers only include hard handovers. All UEs are admitted to a specific
target cell after disconnecting serving cell communication links.
Handover Trigger Causes
Network coverage
trigger
Network load
trigger
Speed trigger
Service trigger
The quality of neighboring cell signals is superior than that of serving cell signals.
Moreover, the servicing cell signal quality falls below a specific threshold.
Serving cells are overloaded. Neighboring cells bear few loads. The neighboring cell
quality detected by the UE satisfies a specific threshold.
A specific UE-initiated service is not supported by serving cells but supported by
neighboring cells. The neighboring cell quality detected by the UE satisfies a specific
threshold.
When the eNodeB judges that the UE moves at a speed higher or lower than a specific
speed and the network where the UE camps on is deployed with high-speed and lowspeed cells, the eNodeB hands over the UE to a corresponding cell.
Intra-eNodeB Handover Success Rate

Intra-eNodeB handovers contain intra-frequency and inter-frequency handovers,
both supporting incoming and outgoing handovers and requiring separate
calculations of handover success rates.
 Success rate of intra-eNodeB intra-frequency incoming/outgoing handovers ＝
Number of successful intra-eNodeB intra-frequency incoming/outgoing
handovers/Number of requests for intra-eNodeB intra-frequency
incoming/outgoing handovers
 Success rate of intra-eNodeB inter-frequency incoming/outgoing handovers ＝
Number of successful intra-eNodeB inter-frequency incoming/outgoing
handovers/Number of requests for intra-eNodeB inter-frequency
incoming/outgoing handovers  100%
UE
eNB
RRC Reconfiguretion/
Handover Command
RRC Connection Setup Procedure
RRC Connection Reconfiguration /
Handover Confirm
Measurement-based and non-measurementbased handovers must be considered when you
calculate handover counts.
(1) Non-measurement-based handovers are
triggered by judging load balance or more.
(2) Measurement-based handovers are triggered by
delivering measurement reports.
X2 Handover Success Rate 1


This index reflects the success rate for a UE connected
with multiple eNodeBs over the X2 interface to hand over
from one eNodeB to another. It relates to the system
handover processing capability and network optimization
and is an important index affecting direct user experience.
X2 handovers contain intra-frequency and inter-frequency
handovers, both supporting incoming and outgoing
handovers and requiring separate calculations of handover
success rate.


Success rate of intra-frequency X2 incoming/outgoing handovers＝
(Number of successful intra-frequency X2 incoming/outgoing
handovers/Number of requests for intra-frequency X2
incoming/outgoing handovers)  100%
Success rate of inter-frequency X2 incoming/outgoing handovers＝
(Number of successful inter-frequency X2 incoming/outgoing
handovers/Number of requests for inter-frequency X2
incoming/outgoing handovers)  100%
X2 Handover Procedure and Signaling Point
Statistics
Outgoing
UE
Source
eNB
Target
eNB
MME
SGW
PDN GW
handover request
(1) Measurement Report
(2) Handover Request
(3) Handover Request Ack
(4) RRC Reconfiguretion/
Handover Command
Outgoing handover
completed
(5) SN Status Transfer
(6) Data Forwarding
Incoming handover
request
(7) RRC Connection Setup Procedure
(8) RRC Connection Reconfiguration
/ Handover Confirm
(9) Path Switch Request
(10) Create Bearer Request
(11) Update Bearer Request
(12) Update Bearer Response
(13) Create Bearer Response
(14) Path Switch Request Ack
(15) End Marker
(16) End Marker
(17) UE Context Release
Incoming handover
completed
S1 Handover Success Rate


When the eNodeB decides to perform a UE handover based on UE
measurement results and the target cell is not connected to this
eNodeB over an X2 interface, hand over the UE over an S1 interface
on the core network. This index relates to the system handover
processing capability and network planning and is an important index
affecting direct user experience.
S1 handovers contain intra-frequency and inter-frequency handovers,
both supporting incoming and outgoing handovers and requiring
separate calculations.


Success rate of intra-frequency S1 incoming/outgoing handovers ＝
(Number of successful intra-frequency S1 incoming/outgoing
handovers/Number of requests for intra-frequency S1 incoming/outgoing
handovers) × 100%
Success rate of inter-frequency S1 incoming/outgoing handovers ＝
(Number of successful inter-frequency S1 incoming/outgoing
handovers/Number of requests for inter-frequency S1 incoming/outgoing
handovers) × 100%
S1 Handover Procedure and Signaling Point
Statistics
UE
Source
eNB
Target
eNB
Source
MME
Target
MME
Source
SGW
Target
SGW
PDN GW
(1) Measurement Report
(2) Handover Required
(3) Forward Relocation Request
(4) Create Bearer Request
(5) Create Bearer Response
(6) Handover Request
(7) Handover Request Ack
(8) Create Bearer Request
(9) Create Bearer Response
(10) Forward Relocation Response
(11) Create Bearer Request
(12) Create Bearer Response
(13) Handover Command
(14) RRC Reconfiguretion/
Handover Command
(15) eNB Status Transfer
(16) Forward SRNS Context
(17) Forward SRNS Context Ack
(18) eNB Status Transfer
(19) Data Forwarding
(20) RRC Connection Setup Procedure
(21) RRC Connection Reconfiguration
/ Handover Confirm
(22) Handover Notify
(23) Forward Relocation Complete
(24) Forward Relocation Complete Ack
(25) Update Bearer Request
(26) Update Bearer Request
(27) Update Bearer Response
(28) Update Bearer Response
Inter-System Handover Success Rate

TD-LTE system may coexist with GSM, WCDMA, CDMA, and TD-SCDMA
systems as a result of network construction. The handover from one system to
another is called an inter-system handover.
 Incoming and outgoing handovers must be calculated. Inter-system handover
success rates are given in a similar formula. The following equations take the
cell handover between TD-LTE and CDMA as an example to calculate the
success rate.
 Inter-system cell handover success rate (LTE-> CDMA) = Success rate of
inter-system packet domain outgoing handovers (EPS-> CDMA) = (1 –
Number of unsuccessful inter-system packet domain outgoing handovers
(EPS-> CDMA)/Number of requests for inter-system packet domain
outgoing handovers (EPS-> CDMA)) × 100％
 Inter-system cell handover success rate (CDMA ->LTE) = Success rate of
inter-system packet domain incoming handovers (CDMA ->EPS) =
(Number of successful inter-system packet domain incoming handovers
(CDMA ->EPS)/Number of requests for inter-system packet domain
incoming handovers (CDMA ->EPS)) × 100％
Handover From LTE to Other Wireless Networks
UE
Source
eNB
Sourse
MME
Target RAN(W/C/T)
CN&AN
(1) Measurement Report
(2) Handover Required
(3) Forward Relocation Request
(4) Other RAN Preparation
(5) Forward Relocation Response
(6) Handover Command
(7) RRC Reconfiguretion/
Handover Command
(8) Handover to target RAN
(9) Forward Relocation Complete
(10) Release Resourse

Inter-system
packet domain
outgoing
handover request
Handover From Other Wireless Networks to LTE
UE
Source
eNB
Sourse
MME
Target RAN(W/C/T)
CN&AN
(1) Handover initiation
(2) Forward Relocation Request
(3) Handover Request
(4) Allocation Resource
(5) Handover Request Ack
(6) Forward Relocation Response
(7) Handover command from other RAN
(8) RRC Connection Setup Procedure
(9) RRC Connection Reconfiguration
/ Handover Confirm
(10) Handover Notify
(11) Forward Relocation Complete

Inter-system
packet domain
incoming
handover request
 Handover
completed
Contents





KPI Overview
KPI Establishment
KPI Report Procedure
Index Classification
Key Index Analysis






Coverage Indexes
Call Setup Indexes
Call Retention Indexes
Mobility Management Indexes
Quality Indexes
System Resource Indexes
Uplink/Downlink Block Error Rate



PUSCH (PDSCH) block error rates are an important index
reflecting the wireless interface signal transmission quality,
the basis for managing and controlling wireless resources,
and a factor affecting system handover, power control
admission, and other aspects. This index indicates network
coverage conditions, reflects networking interference
conditions, and is an index indirectly reflecting the network
planning quality and algorithm-related quality.
Uplink block error rate ＝ (Number of CRC errors of
received uplink transmission blocks/Number of received
uplink transmission blocks in total) × 100%
Downlink block error rate ＝ (Number of CRC errors of
received downlink transmission blocks/Number of received
downlink transmission blocks in total) × 100%
Uplink/Downlink MAC Layer Retransmission Rate

Retransmission rate is a key ind reflecting the packet
service quality and the HARQ performance. Improving
retransmission rates is an important way to optimize
packet service networks.
 You can calculate the uplink MAC layer retransmission
rate by obtaining the confirmed number of MAC PDUs
received by the eNodeB and the total number of received
MAC PDUs.
 Uplink MAC layer retransmission rate = 1- Confirmed
number of PDUs received in the uplink MAC layer/Number
of PDUs received in the uplink MAC layer
 Downlink MAC layer retransmission rate = 1- Confirmed
number of PDUs received in the downlink MAC
layer/Number of PDUs received at the downlink MAC layer
Contents





KPI Overview
KPI Establishment
KPI Report Procedure
Index Classification
Key Index Analysis






Coverage Indexes
Call Setup Indexes
Call Retention Indexes
Mobility Management Indexes
Quality Indexes
System Resource Indexes
Traffic Index—S1 Interface Traffic




This index reflects system load conditions at S1s and
facilities overall understanding of service usage conditions.
It refers to the traffic calculated at the transport layer (IP
layer) over S1s. As the uplink and downlink traffic volumes
are symmetric, S1 traffic supports two indexes: S1 uplink
(in relation to eNodeB) traffic and S1 downlink (in relation
to eNodeB) traffic.
S1 uplink traffic ＝ Sum of data rates sent by all S1s and
IP addresses to the IP layer
S1 downlink traffic ＝ Sum of data rates received by all
S1s and IP addresses from the IP layer
Unit: bit/s
Traffic Index—X2 Traffic (Temporarily Unavailable)

This index reflects system load conditions at X2s
and refers to the traffic calculated the transport
layer (IP layer) over X2s. As the uplink and
downlink traffic volumes are symmetric, X2 traffic
supports two indexes: X2 uplink (in relation to
eNodeB) traffic and X2 downlink (in relation to
eNodeB) traffic.
 X2 uplink traffic ＝ Sum of data rates sent by all
X2s and IP addresses to the IP layer
 X2 downlink traffic ＝ Sum of data rates received
by all X2s and IP addresses from the IP layer
 Unit: bit/s
Traffic Index—Traffic at the MAC Layer








This index reflects system load conditions at the MAC layer and
indicates network load conditions as well as system processing
capability to some extent. It is separately calculated based on uplink
(UE--->eNodeB) and downlink (eNodeB-->UE) directions.
Cell uplink traffic at the MAC layer ＝ Data rates received at the MAC
layer
Cell downlink traffic at the MAC layer ＝ Data rates sent from the MAC
layer
Data rates received at the MAC layer ＝ Number of all PDUs received
by the eNodeB in the uplink MAC layer from the UE
Data rates sent from the MAC layer ＝ Number of all PDUs sent by the
eNodeB at the downlink MAC layer
eNodeB uplink flow at the MAC layer ＝ Sum of data rates received by
all eNodeB cells at the MAC layer (temporarily unavailable)
eNodeB downlink flow at the MAC layer ＝ Sum of data rates sent by
all eNodeB cells at the MAC layer (temporarily unavailable)
Unit: Mbit/s
Traffic Index—Traffic at the PDCP Layer






This index reflects system load conditions at the PDCP layer and
indicates network load conditions as well as system processing
capability to some extent. It is separately calculated based on uplink
(UE--->eNodeB) and downlink (eNodeB-->UE) directions.
Cell uplink traffic (all QCIs) at the PDCP layer ＝ Data rates received
at the cell PDCP layer
Cell downlink traffic (all QCIs) at the PDCP layer ＝ Data rates sent
from the cell PDCP layer
eNodeB uplink flow at the PDCP layer ＝ Sum of data rates received
by all eNodeB cells at the PDCP layer (temporarily unavailable)
eNodeB downlink flow at the PDCP layer ＝ Sum of data rates sent by
all eNodeB cells at the PDCP layer (temporarily unavailable)
Unit: Mbit/s
Average PUSCH PRB Utilization



This index reflects radio resource utilization (in relation to
PRBs assigned to the PUSCH) and provides the basis for
optimizing system algorithms and deciding whether the
system requires expansion.
Average PUSH PRB utilization = Average value of
utilization of all TTI PUSCH PRBs in a statistical cycle =
Number of carrier-occupied PUSCH PRBs/Total number of
cell-occupied PRBs in the uplink
TTI PUSCH PRB utilization = Number of PUSCH PRBs
occupied per TTI/ Total number of PUSCH PRBs per TTI
Average PDSCH PRB Utilization

This index indicates the utilization of PRBs
assigned to the PDSCH.
 Average PDSH PRB utilization = Average value of
utilization of all TTI PDSCH PRBs in a statistical
cycle = Number of cell-occupied PRBs in the
downlink/Total number of cell-occupied PRBs in
the uplink
 TTI PDSCH PRB utilization = Number of PDSCH
PRBs occupied per TTI/ Total number of PDSCH
PRBs per TTI
Average PRACH Resource Utilization





This index reflects the system access capability and
provides the basis for optimizing system parameters to
some extent.
Average PRACH resource utilization ＝ Number of PRACH
resources in use/Number of configured PRACH resources
Namely, = Number of carrier-occupied PRACHs/Number of
cell-occupied PRBs in the uplink
PRACH resources are a collection of time domains,
frequency domains, and code domains.
The number of PRACH resources in use is the number of
PRACHs which the eNodeB successfully resolves their
preamble, including public and dedicated preambles.
Maximum/Average Carrier Transmit Power
Utilization (Temporarily Unavailable)

By statistically obtaining the maximum and average carrier transmit
power values allows you to estimate the maximum and average
downlink system loads within a specific period, thereby knowing the
carrier transmit power utilization within this period. During network
planning, the system implements link budge calculation and capacity
emulation according to preset downlink loads. Therefore, to calculate
downlink carrier transmit utilization of each network cells has great
significance in network expansion planning and network optimization.

Maximum carrier transmit power utilization ＝ Maximum carrier
transmit power/Configured maximum carrier transmit power × 100%

Average carrier transmit power utilization ＝ Average carrier transmit
power/Configured maximum carrier transmit power × 100%
Paging Congestion Rate





This user-perceptible index indirectly reflects radio paging
channel utilization. It relates with paging message delivery
failures due to eNodeB resource limitations.
Paging congestion rate ＝ Number of paging
congestions/Number of paging attempts × 100%
Where, Number of paging attempts ＝ Number of paging
congestions ＋ Number of successful paging attempts
(1) Number of paging congestions: refers to the number of
paging failures. The eNodeB sends a paging message to
the UE. The FP fails to add a paging record to the PCCH
frame because paging occasions are full.
(2) Number of successful paging attempts: The eNodeB
receives the RRCConnectionSetupComplete message
from the UE in response to the paging message.