L21Q3KPI

Key Performance Indicators
User Guide
37/1553-LZA 701 6014/1 Uen H

Copyright
© Ericsson AB 2021. All rights reserved. No part of this document may be

reproduced in any form without the written permission of the copyright owner.
Disclaimer
The contents of this document are subject to revision without notice due to
continued progress in methodology, design and manufacturing. Ericsson shall
have no liability for any error or damage of any kind resulting from the use of this
document.
Trademark List
37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Contents
Contents
1 Introduction 1
1.1 End-User Performance in LTE 1
1.2 Observability in Ericsson LTE 2
1.3 LTE RAN Performance Observability Model 3
1.4 Aggregation of KPIs 4
1.5 Service Separation of Observability in LTE RAN 6
1.6 Detailed Separation of Observability in LTE RAN with Flexible
Counters 7
2 Accessibility 9
2.1 Initial E-RAB Establishment Success Rate 9
2.2 Initial E-RAB Establishment Success Rate per QCI 19.Q4 and
later 10
2.3 Differentiated Initial E-RAB Establishment Success Rate 13
2.4 Compensated Initial E-RAB Establishment Success Rate 15
2.5 Initial E-RAB Establishment Success Rate for Emergency Calls 17
2.6 Initial E-RAB Establishment Success Rate, No MO Signaling 18
2.7 Differentiated Initial E-RAB Establishment Success Rate, No
MO Signaling 20
2.8 Added E-RAB Establishment Success Rate 22
2.9 Differentiated Added E-RAB Establishment Success Rate 23
2.10 Compensated Added E-RAB Establishment Success Rate 24
2.11 Added E-RAB Establishment Success Rate per QCI 25
2.12 Added E-RAB Establishment Success Rate for Emergency Calls 26
2.13 Robust Random Access Success Rate 27
2.14 Random Access MSG1 Success Rate 27
2.15 Robust Random Access Success Rate for Cat-M1 UE devices 28
2.16 Random Access MSG1 Success Rate for Cat-M1 UE devices 29
2.17 eMBMS Establishment Success Rate 29
3 Retainability 31
3.1 E-RAB Retainability - Session Time Normalized Loss Rate 31
3.2 E-RAB Retainability - Session Time Normalized Loss Rate the
Second 32
3.3 Compensated E-RAB Retainability - Session Time Normalized
Loss Rate 33
37/1553-LZA 701 6014/1 Uen H | 2021-08-25

3.4 Compensated E-RAB Retainability - Session Time Normalized

Loss Rate, the Second 35
3.5 E-RAB Retainability - Session Time Normalized per QCI Loss
Rate 36
3.6 E-RAB Retainability - Session Time Normalized per QCI Loss
Rate, the Second 37
3.7 E-RAB Retainability - Percentage Lost 39
3.8 Differentiated E-RAB Retainability - Percentage Lost 40
3.9 Compensated E-RAB Retainability - Percentage Lost 41
3.10 E-RAB Retainability (eNB) - Percentage Lost 42
3.11 E-RAB Retainability (eNB) - Percentage Lost, the Second 43
3.12 Compensated E-RAB Retainability (eNB) - Percentage Lost 45
3.13 Compensated E-RAB Retainability (eNB) - Percentage Lost,
the Second 46
3.14 E-RAB Retainability - Percentage Lost per QCI 48
3.15 E-RAB Retainability - Percentage Lost for Emergency Calls 49
3.16 eMBMS Retainability - Percentage Lost 50
4 Integrity 52
4.1 Average DL UE Latency 52
4.2 Compensated Average DL UE Latency 52
4.3 Average DL PDCP DRB LTE Latency per QCI 53
4.4 Average DL PDCP DRB Cat-M Latency per QCI 54
4.5 Average DL UE PDCP DRB Latency per QCI 55
4.6 Average DL MAC DRB Latency per QCI 56
4.7 Differentiated Average DL MAC DRB Latency per QCI 57
4.8 Compensated average DL MAC DRB Latency per QCI 58
4.9 Average DL PDCP UE Throughput 59
4.10 Compensated Average DL PDCP UE Throughput 60
4.11 Total DL PDCP UE Throughput 61
4.12 Average DL PDCP Cell Throughput 62
4.13 Average DL MAC Cell Throughput 63
4.14 Filtered MBB UE DL PDCP Throughput 64
4.15 DL PDCP UE Throughput Distribution 65
4.16 DL PDCP UE Throughput Distribution for MBB UE 65
4.17 Mean DL PDCP UE Throughput for MBB UE 66
4.18 Mean DL PDCP UE Throughput 67
4.19 Average DL PDCP UE Throughput for Carrier Aggregation 68
37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Contents
4.20 Compensated Average DL PDCP UE Throughput for Carrier

Aggregation 69
4.21 Filtered MBB UE UL Throughput 70
4.22 Average DL PDCP UE DRB Throughput per QCI 71
4.23 Average UL PDCP UE Throughput 72
4.24 Compensated Average UL PDCP UE Throughput 74
4.25 Average UL PDCP Cell Throughput 74
4.26 Average UL MAC Cell Throughput 75
4.27 Normalized Average UL MAC Cell Throughput Considering
Successful PUSCH subframe Only 76
4.28 Normalized Average UL MAC Cell Throughput Considering
Successful PUSCH Slot Only 77
4.29 Average UL PDCP UE Throughput for Carrier Aggregation 78
4.30 Compensated Average UL PDCP UE Throughput for Carrier
Aggregation 79
4.31 Average MAC UE UL THROUGHPUT 80
4.32 UL PDCP UE Throughput Distribution 80
4.33 Mean UL PDCP UE Throughput 81
4.34 Average UL UE Throughput per LCG 81
4.35 Compensated average UL Throughput per LCG 82
4.36 Average DL Packet Error Loss Rate 83
4.37 Average DL Packet Error Loss Rate per QCI 85
4.38 Average eMBMS Packet Loss Rate 87
4.39 Average UL Packet Loss Rate 88
4.40 Average UL Packet Loss Rate per QCI 89
4.41 VoIP Cell Integrity 90
4.42 Compensated VoIP Cell integrity 91
4.43 GTP-U Downlink Packet Loss Rate 92
4.44 GTP-U Downlink Packet Out of Order Rate 92
4.45 DL Scheduled Entities per TTI 93
4.46 UL Scheduled Entities per TTI 94
5 Mobility 95
5.1 Mobility Success Rate 95
5.2 Cell Mobility Success Rate 97
5.3 Flexible Intra-Frequency Cell Mobility Success Rate 99
5.4 Handover Success Rate 101
5.5 Handover Execution Success Rate 103
5.6 Cell Handover Success Rate 104
37/1553-LZA 701 6014/1 Uen H | 2021-08-25

5.7 Cell Handover Execution Success Rate 106

5.8 Differentiated Cell Handover Success Rate 108
5.9 UTRAN SRVCC Success Rate 110
6 Availability 112
6.1 Cell Availability 112
37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Introduction
1 Introduction
KPIs are used for RAN to measure the quality and system performance observed
and perceived by subscribers.
Monitoring the performance of a RAN through KPIs enables the following

activities:
— Supervising and optimizing radio network performance to provide better
quality or better use of available resources.
— Detecting performance issues in the network to preserve service quality.
— Providing radio network planners with detailed information for

dimensioning, optimizing, and configuring the network.
— Troubleshooting on cell clusters of interest.
Performance Management (PM) counters collect performance statistics from

measurements on live traffic in the radio network. These measurements do not
interfere with user connection and transmission handling. The nodes can
sometimes delay data collection and postpone the measurement period due to
increased resource use. In this case, the results of the KPI calculations can be
observed as a minor peak in the overall performance statistics.
1.1 End-User Performance in LTE

Key Performance Indicators (KPIs) in LTE networks measure both end-user
performance and network performance in the RAN. The performance of end-user
applications is measured over a broader area than the RAN. As illustrated on , it
covers the entire packet delivery chain.
Application UE RBS TN SGW TN PGW Server

layer
L0000266A
PGW Packet Data Network Gateway
SGW Serving Gateway
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 1

TN Transport Network
UE User Equipment
Figure 1 LTE Packet Delivery Chain
1.2 Observability in Ericsson LTE

Observability covers all functions in LTE RAN that supervise the performance and
characteristics of the network. This can be done on various levels with different
target groups and requirements. Figure 2 illustrates a model for observability in
LTE RAN.
Key Performance Indicator (KPI) level End-user

perception
Performance Indicator (PI) level
Procedure level
System
Characteristics
L0000269A
Figure 2 Top-Down Approach in LTE RAN Observability
Table 1 Observability Levels and Purposes

Observability Level Description of Use
Key Performance Indicator (KPI) The KPI represents the end-user
perception of a network on a macro
level. Operators use KPI statistics to
compare networks against each other,
and to detect problems and errors.
KPIs are calculated from PI and PM
counters. The reliability, granularity,
and accuracy of the data are critical,
and the data is collected continuously.
Performance Indicator (PI) The PI represents system level
information that explains the KPI
results. Many PIs can be based on PM
counters, for example, PM counters for
Root Cause Analysis. PIs can also be
2 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Introduction
Observability Level Description of Use

in the form of metrics that show
specific parts of the system can
perform. PIs usually have an impact
on KPIs. PIs can be collected on
demand.
RI The RI represents system level
information that explains the system
capacity and resource use. The RI
data can be used for planning and
dimensioning. RI data, mostly from
PM counters, can be collected
continuously.
Procedure The Procedure level represents in-
depth troubleshooting, and the
measurements of system
characteristics. Signalling, procedure,
and function levels are included in the
measurements. The purpose of the
measurements is to investigate
problems detected at higher levels.
The measurements are generally user-
initiated for a specific purpose or for a
specific area. The scope of the
measurements are limited with the
specified aspects. The source for the
data is mostly the UE trace and the
cell trace recording functions.
For more information about UE trace and cell trace, see the document
Performance Management (DU node) or Manage Performance (Baseband node).
1.3 LTE RAN Performance Observability Model

The LTE RAN Performance Observability model used by Ericsson combines the
general, International Telecommunications Union-Telecommunications (ITU-T)
QoS categories with mobility, availability and utilization.
Table 2 Ericsson Observability Model

QoS Categories Description
Accessibility The ability of a service to be obtained,
within specified tolerances and other
given conditions, when requested by
the end user.
Retainability The probability that a service, once
obtained, remains provided under
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 3

QoS Categories Description

given conditions for a given time
duration.
Integrity The degree to which a service is
provided without excessive
impairments, once obtained. Service
Integrity represents the quality
experienced by the end user during the
call or session.
Mobility The ability of the system to allow
movement within the LTE RAN.
Availability The ability of an item to be in a state
to perform a chosen function at a
given instant of time within a given
time interval, assuming that the
external resources, if necessary, are
provided.
Utilization Describes the network use by
measuring traffic level and capacity
resource management, including
congestion, admission and load
control, and license use. Utilization
information is needed as input to
network planning. For formulas to
calculate utilization, see License and
Resource Use Indicators.
In LTE, all sessions are Packet Switched (PS). The LTE network supports multiple
simultaneous services per user. ITU-T definitions are used as guidelines for
performance measurement.
KPIs are developed for observing the network performance impact on the end
user, and for observing the performance of the network itself. Each KPI is defined
for observing end-user impact or system performance.
1.4 Aggregation of KPIs

The PM counters can be aggregated in the following ways:
— Over a group of cells
— Over a cluster of cells
— Over one RBS
— Over multiple RBSs
— Over a complete geographic area
4 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Introduction
— Over time periods (hour, days, week)
Aggregation needs to be considered for maximum calculations in case of some

PIs, where the maximum value means the average of maximum or the maximum
of average. A distribution function can be created as an alternative for the set of
max values, so that the range of values can be preserved and highlighted. In
some cases, the data-capture tool and also the user data can provide some
further aggregation.
Many KPIs are basically a ratio of two quantities (for example, between success
and failure, or between volume and time). Based on the two quantities, two basic
aggregation methodologies are available that can produce different results. The
operator needs to consider exactly what is calculated, and also know the
advantage and disadvantage of the calculated quantities.
The PM counters can be aggregated over time (for example, hour, day or week)
for a specific cell to get a cell-specific, time aggregated PI. The cell PI can be
aggregated over cell clusters to get an average over an area. Use both equations
for the aggregation.
Example of Aggregation
1. Aggregate the individual PM counter values across the time and geographic
area using the equation.
2. Perform the PI equation.
In certain situations, an operator needs to replace the individual pmXyz PM

counter in each equation to get a metric over a larger area and longer time, or for
both.
1. Calculate the PI equation for the specific metric at the lower time and
geographic resolution with the formula.
2. Aggregate across the longer time and larger geographic area using the
equation.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 5

Note: When a formula involves the sum of fractions (for example, the sum of
PI(Cell&ROP), the result of the formula indicates an invalid result that
is invalid for all fractions involved. When the result is invalid for less than
all fractions involved in the formula, the invalid fractions (for example,
0 / 0) are ignored. In this case, the value of " N " is to be adjusted.
Result of Aggregation
— Use of aggregated pmXyz can mask individual cells or times when there are
not usual results in a specific time or location. This can result in an overall
average.
— Use of PI(Cell&ROP) can highlight the impact of individual metrics as each

PI is equally weighted. This results in equal amount of weight for cells with
minimal traffic and with large traffic. The final result can be biased.
— In cases where the denominator pmRst(Cell&ROP) is a fixed number in the

samples (for example, 900 seconds in a ROP), then both methods produce
the same result.
— In case of throughput metrics, SUM(all_Volume) / SUM(all_Time), can

introduce significant bias. For example, if a few UEs require a long delivery
time because of poor RF. This can hide the benefits of features such as
Carrier Aggregation and Frequency Selective Scheduling that attempt to
maximize individual UE throughput.
— Aggregation of quantities based on time (for example, pmUeThpTimeDl and

pmDrbThpTimeDlQci time counters, or uplink and downlink throughput
metrics) need special consideration as aggregation when metric occurs
simultaneously on same entity. Results can differ significantly from
aggregation when there is no simultaneity. In reality, there is often a mixture
of simultaneous and sequential actions which are challenging to interpret.
Simultaneity can impact the definition of some PM counters. The
pmPdcpVolDlDrbLastTti and SUM(pmPdcpVolDlLastTtiQci) PM Counters
can be different depending on the simultaneity (at TTI level) of the individual
QCI bursts.
— Same aggregation methodology needs to be used comparing results to avoid

any differences triggered by the aggregation.
— SUM function needs to be avoided, if the aggregated quantities are

depending on each other. For example, the volume and time metrics related
to lastTt.
— The type of aggregation can be decided based on the date capture tool or the
context and purpose of use.
Aggregation is indicated in the formulas when in use.

1.5 Service Separation of Observability in LTE RAN
In an LTE system, the service is not directly known to the RAN. If a service is
mapped towards a specific Quality of Service Class Indicator (QCI) or Allocation
6 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Introduction
Retention Priority (ARP), then it is possible to determine the observability of the

service.
Emergency Calls If it is defined that Emergency (EM) Calls have a specific
AR, accessibility for Emergency calls can be determined
by using the Initial E-RAB Establishment Success Rate
for Emergency Calls formula.
Web sessions If seven is defined as QCI for Web sessions, downlink

Throughput for Web sessions can be determined using
the DL Throughput per QCI formula.
VoIP If one is defined as QCI for VoIP, Retainability for VoIP

can be determined using the E-RAB Retainability -
Percentage Lost per QCI formula.
Note: The ARP and QCI values are only examples. More information about the
different uses of ARP and QCI can be found in TS 23.203.
For the uplink, the scheduler grants resources based on Logical Channel Groups
(LCG) instead of QCI. Some uplink measurement per LCG is preferable than per
QCI. For more information about LCG and QCI, see Quality of Service (DU node)
or Manage Quality of Service (Baseband node).
1.6 Detailed Separation of Observability in LTE RAN with

Flexible Counters
When new UE categories are introduced with different characteristics, for
example Cat-M1 UE devices, KPIs can show higher or lower values. The solution
can be using the Flexible Counters feature or using PM counters that are related
to the specific characteristics.
The Flexible Counters feature can be used for the following purposes:
— Replacing the existing PM counters with the use of flexible counters and
filters to the related PM counters. For example, if two PLMNs are used and
the result for throughput is needed separately. The flexible counters filtered
on the two PLMNs replace the legacy PM counters in the corresponding KPI.
— Compensating KPIs with subtracting legacy PM counters of a KPI formula

and replacing it with flexible counters filtered on a specific characteristic. For
example, Cat-M1 UE part is subtracting from the throughput KPI. The
maximum throughput for Cat-M1 UE devices is lower than the average,
which results in overall lower average for UE throughput. This can be
compensated by using flexible counters filtered on the Cat-M1 UE category.
Each KPI formula that can be used with flexible counters are indicated with an
example.
Note: The Flexible Counters feature is not supported on DUL 20 and Pico RBS.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 7

For list of filters supported by flexible counters, see Flexible Counters.
Note: These figures are only examples of KPIs, and cannot be used for
measurement purposes.
8 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
2 Accessibility
2.1 Initial E-RAB Establishment Success Rate

This KPI measures the impact on the end user.
The following equation gives the accessibility success rate for end-user services
that are carried by E-RABs included in the Initial UE Context setup procedure.
The PM counters in this KPI are on cell level.
Network Function
LTE
Level EUtranCell
Formula Status CURRENT
Calculate Initial E-RAB Establishment Success Rate by using the following

formula:
The results of the following subformulas are used in the calculation of Initial E-
RAB Establishment Success Rate:
Figure 3 Initial E-RAB Establishment Success Rate
The following counters are used in the formulas:
— pmRrcConnEstabSucc
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 9

— pmRrcConnEstabAtt
— pmRrcConnEstabAttReatt
— pmRrcConnEstabFailMmeOvlMos
— pmRrcConnEstabFailMmeOvlMod
— pmS1SigConnEstabSucc
— pmS1SigConnEstabAtt
— pmS1SigConnEstabFailMmeOvlMos
— pmErabEstabSuccInit
— pmErabEstabAttInit
Note: — The impact of multiple E-RABs in the Initial context setup procedure
is not considered in the RRC and S1 signaling part contribution.
Compared to 3GPP 32.450 definition of the KPI, Ericsson subtract
repetitive attempts from the same UE, see definition of
pmRrcConnEstabAttReatt.
— During the high load or overload an RRC Connection Establishment

reattempt can be rejected. In such a case PM counter
pmRrcConnEstabAtt is pegged, PM counter
pmRrcConnEstabAttReAtt is NOT pegged for the reattempt. Double
counting of the attempt is possible.
— The PM counter pmRrcConnEstabAtt is not stepped when the cell

soft lock has started, even though the UE can still try to make RRC
Connection Requests until the cell becomes locked and UE receives
an RRC connection reject
— These KPIs are differentiated using flexible counters see

Differentiated Initial E-RAB Establishment Success Rate.
— Cat-M1 UE devices impact in this KPI can be compensated and the

corresponding Cat-M1 only KPI is also available see Compensated
Initial E-RAB Establishment Success Rate.
2.2 Initial E-RAB Establishment Success Rate per QCI 19.Q4

and later
The PM counters in this KPI are on cell level and per QCI.
10 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
Network Function
LTE
Level EUtranCell
Formula Status PRELIMINARY
Calculate Initial E-RAB Establishment Success Rate per QCI 19.Q4 and later by
using the following formula:
RAB Establishment Success Rate per QCI 19.Q4 and later:
Figure 4 Initial E-RAB Establishment Success Rate per QCI 19.Q4 and later
— pmRrcConnEstabSuccMod
— pmRrcConnEstabSuccMoVoice
— pmRrcConnEstabSuccMta
— pmRrcConnEstabSuccHpa
— pmRrcConnEstabAttMod
— pmRrcConnEstabAttMoVoice
— pmRrcConnEstabAttMta
— pmRrcConnEstabAttHpa
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 11

— pmRrcConnEstabAttReattMod
— pmRrcConnEstabAttReattMoVoice
— pmRrcConnEstabAttReattMta
— pmRrcConnEstabAttReattHpa
— pmS1SigConnEstabSuccMod
— pmS1SigConnEstabSuccMoVoice
— pmS1SigConnEstabSuccMta
— pmS1SigConnEstabSuccHpa
— pmS1SigConnEstabAttMod
— pmS1SigConnEstabAttMoVoice
— pmS1SigConnEstabAttMta
— pmS1SigConnEstabAttHpa
— pmErabEstabSuccInitQci
— pmErabEstabAttInitQci
12 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
pmRrcConnEstabAttReattMod and pmRrcConnEstabAttReattMta.
If the pm...Mod and pm...Mta sub-components are not available,
then the parent PM counters can be used.

reattempt can be rejected, although pmRrcConnEstabAttxxx are
pegged, pmRrcConnEstabAttReAttxxx are NOT pegged for the
reattempt. Double counting of the attempt is possible.
— PM counter pmRrcConnEstabAttMod is not stepped when the cell

Connection Requests until the cell becomes locked and UE receive
an RRC connection reject.
— Several of these PM counters are not included in any predefined

scanner.
— See Detailed separation of Observability in LTE RAN for related

references to the Feature Description of Category M Access.
— Filter parameters can be used for flexible counters. See Detailed

separation of Observability in LTE RAN for more information. These
flexible counters are not included in any predefined scanner.
2.3 Differentiated Initial E-RAB Establishment Success Rate

Network Function
LTE
Level EUtranCell
Calculate Differentiated Initial E-RAB Establishment Success Rate by using the

following formula:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 13

The results of the following subformulas are used in the calculation of

Differentiated Initial E-RAB Establishment Success Rate:
Figure 5 Differentiated Initial E-RAB Establishment Success Rate
— pmFlexErabEstabSuccInit
— pmFlexErabEstabAttInit
14 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility


— These KPIs are differentiated using flexible counters.

Initial E-RAB Establishment Success Rate.

2.4 Compensated Initial E-RAB Establishment Success Rate

Network Function
LTE
Level EUtranCell
Calculate Compensated Initial E-RAB Establishment Success Rate by using the

following formula:

Compensated Initial E-RAB Establishment Success Rate:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 15

Figure 6 Compensated Initial E-RAB Establishment Success Rate
16 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility



Differentiated Initial E-RAB Establishment Success Rate.
— Please see Detailed separation of Observability in LTE RAN for

related references to the Feature Description of Category M Access.

2.5 Initial E-RAB Establishment Success Rate for

Emergency Calls
The following equation gives the accessibility success rate for emergency calls
The PM counters in this KPI are on cell level and per ARP (hence only include the
ARP range that defines Emergency).
Network Function
LTE
Level EUtranCell
Calculate Initial E-RAB Establishment Success Rate for Emergency Calls by using
the following formula:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 17

Figure 7 Initial E-RAB Establishment Success Rate for Emergency Calls
— pmRrcConnEstabSuccEm
— pmRrcConnEstabAttEm
— pmRrcConnEstabAttReattEm
— pmS1SigConnEstabSuccEm
— pmS1SigConnEstabAttEm
— pmErabEstabSuccInitArp
— pmErabEstabAttInitArp
Note: — If an RRC Connection Establishment reattempt is rejected because

of high load or overload, then although pmRrcConnEstabAttEm are
pegged, pmRrcConnEstabAttReAttEm are NOT pegged for the
reattempt, thus some double counting of the attempt is possible.
— The PM counter pmRrcConnEstabAttEm is not stepped when the

cell soft lock has started, even though the UE can still try to make
RRC Connection Requests until the cell becomes locked and UE
receives an RRC connection reject.

scanner
2.6 Initial E-RAB Establishment Success Rate, No MO

Signaling
This KPI measures the E-RAB establishment success rate and one can remove all
RRC connection requests that are set up for signaling from the first parts of the
formula. That would result in the following alternative formula for initial E-RAB
establishment success rate, no MO signaling.
Network Function
LTE
Level EUtranCell
18 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
Calculate Initial E-RAB Establishment Success Rate, No MO Signaling by using

RAB Establishment Success Rate, No MO Signaling:
Figure 8 Initial E-RAB Establishment Success Rate, No MO Signaling
— pmRrcConnEstabSuccMos
— pmRrcConnEstabAttMos
— pmRrcConnEstabAttReattMos
— pmS1SigConnEstabSuccMos
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 19

— pmS1SigConnEstabAttMos



Differentiated Initial E-RAB Establishment Success Rate, No MO
Signaling.
2.7 Differentiated Initial E-RAB Establishment Success Rate,

No MO Signaling
This KPI measures the E-RAB establishment success rate and one can remove all
RRC connection requests that are set up for signaling from the first parts of the
formula. That would result in the following alternative formula for initial E-RAB
establishment success rate, no MO signaling.
Network Function
LTE
Level EUtranCell
Calculate Differentiated Initial E-RAB Establishment Success Rate, No MO

Signaling by using the following formula:
20 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility

Differentiated Initial E-RAB Establishment Success Rate, No MO Signaling:
Figure 9 Differentiated Initial E-RAB Establishment Success Rate, No MO

Signaling
— pmRrcConnEstabSuccMos
— pmRrcConnEstabAttMos
— pmRrcConnEstabAttReattMos
— pmS1SigConnEstabSuccMos
— pmS1SigConnEstabAttMos
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 21




2.8 Added E-RAB Establishment Success Rate

This KPI measures the impact on the end user. Accessibility success rate for end-
user services which is carried by E-RABs included in the E-RAB setup procedure is
given by the following equation. The PM counters in this KPI are on cell level. The
KPI excludes the additional E-RAB setup attempts made during ongoing
procedure and setup attempts that lead to CS Fallback.
Network Function
LTE
Level EUtranCell
Calculate Added E-RAB Establishment Success Rate by using the following

formula:
Figure 10 Added E-RAB Establishment Success Rate
— pmErabEstabSuccAdded
22 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
— pmErabEstabAttAdded
— pmErabEstabFailAddedOngoingProc
— pmErabEstabAttAddedCsfb
Note: — This KPI can be compensated or differentiated by using flexible

counters see Compensated added E-RAB Establishment Success
Rate and Differentiated added E-RAB Establishment Success Rate.
2.9 Differentiated Added E-RAB Establishment Success

Rate
procedure and setup attempts that lead to CS Fallback. Formula can be
differentiated with different filter counters when the operator configures them as
pmFlex.
Network Function
LTE
Level EUtranCell
Calculate Differentiated Added E-RAB Establishment Success Rate by using the

following formula:

Differentiated Added E-RAB Establishment Success Rate:
Figure 11 Differentiated Added E-RAB Establishment Success Rate
— pmFlexErabEstabSuccAdded
— pmFlexErabEstabAttAdded
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 23

— pmFlexErabEstabFailAddedOngoingProc
— pmFlexErabEstabAttAddedCsfb
Note: — Filter parameters can be used for flexible counters. See Detailed
— Flexible counters can be filtered for different UE EN-DC stages. See

Detailed separation of Observability in LTE RAN for more
information.
2.10 Compensated Added E-RAB Establishment Success

Rate
procedure and setup attempts that lead to CS Fallback. Formula can be
compensated with different filter counters when the operator configures them as
pmFlex.
Network Function
LTE
Level EUtranCell
Calculate Compensated Added E-RAB Establishment Success Rate by using the

following formula:

Compensated Added E-RAB Establishment Success Rate:
24 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
Figure 12 Compensated Added E-RAB Establishment Success Rate
— pmErabEstabAttAdded
— pmFlexErabEstabAttAdded
— pmErabEstabFailAddedOngoingProc
— pmFlexErabEstabFailAddedOngoingProc
— pmErabEstabSuccAdded
— pmFlexErabEstabSuccAdded
— pmErabEstabAttAddedCsfb
— pmFlexErabEstabAttAddedCsfb

information.
2.11 Added E-RAB Establishment Success Rate per QCI

Accessibility success rate for end-user services which is carried by E-RABs

included in the E-RAB setup procedure is given by the following equation. The
PM counters in this KPI are on cell level per QCI.
The KPI excludes the additional E-RAB setup attempts made during ongoing
handover and setup attempts that lead to CS Fallback.
Network Function
LTE
Level EUtranCell
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 25

Calculate Added E-RAB Establishment Success Rate per QCI by using the
following formula:
Figure 13 Added E-RAB Establishment Success Rate per QCI
— pmErabEstabSuccAddedQci
— pmErabEstabAttAddedQci
— pmErabEstabAttAddedHoOngoingQci
— pmErabEstabAttAddedCsfbQci
Note: — Flexible counters can be used to select measurements for particular

QCIs see Differentiated Added E-RAB Establishment Success Rate.
2.12 Added E-RAB Establishment Success Rate for

Emergency Calls
Accessibility success rate for emergency calls which is carried by E-RABs included
in the E-RAB setup procedure is given by the following equation. The PM
counters in this KPI are on cell level and per ARP (hence only include the ARP
range that defines Emergency).
The KPI excludes the additional E-RAB setup attempts made during ongoing
handover and setup attempts that lead to CS Fallback.
Network Function
LTE
Level EUtranCell
Calculate Added E-RAB Establishment Success Rate for Emergency Calls by

26 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
Figure 14 Added E-RAB Establishment Success Rate for Emergency Calls
— pmErabEstabSuccAddedArp
— pmErabEstabAttAddedArp
— pmErabEstabAttAddedHoOngoingArp
— pmErabEstabAttAddedCsfbArp
2.13 Robust Random Access Success Rate

Robust Random Access Success Rate describes how often a transmitted RaMsg2
results in a successfully received RaMsg3. The PM counters are on a cell level.
Network Function
LTE
Level EUtranCell
Calculate Robust Random Access Success Rate by using the following formula:
Figure 15 Robust Random Access Success Rate
— pmRaSuccCbra
— pmRaMsg2AttCbra
2.14 Random Access MSG1 Success Rate

Random Access MSG1 Success Rate describes how often a detected preamble
results in a successfully allocated RaMsg2 in DL.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 27

Network Function
LTE
Level EUtranCell
Calculate Random Access MSG1 Success Rate by using the following formula:
Figure 16 Random Access MSG1 Success Rate
— pmRaMsg2AttCbra
— pmRaAttCbra
Note: — The PM counters in the Equation are not included in any predefined
scanner.
2.15 Robust Random Access Success Rate for Cat-M1 UE

devices
The following KPI is supported for Cat-M1 UE.
Robust Random Access Success Rate describes how often a transmitted RaMsg2
results in a successfully received RaMsg3.
Network Function
LTE
Level EUtranCell
Calculate Robust Random Access Success Rate for Cat-M1 UE devices by using
Figure 17 Robust Random Access Success Rate for Cat-M1 UE devices
28 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Accessibility
— pmRaSuccCbraBr
— pmRaMsg2AttCbraBr
2.16 Random Access MSG1 Success Rate for Cat-M1 UE

devices
The following KPI is supported for Cat-M1 UE.
Random Access MSG1 Success Rate describes how often a detected preamble
results in a successfully allocated RaMsg2 in DL.
Network Function
LTE
Level EUtranCell
Calculate Random Access MSG1 Success Rate for Cat-M1 UE devices by using
Figure 18 Random Access MSG1 Success Rate for Cat-M1 UE devices
— pmRaMsg2AttCbraBr
— pmRaAttCbraBr
2.17 eMBMS Establishment Success Rate

Accessibility success rate for end-user eMBMS is given by the following equation.
The PM counters in this KPI are on RBS level.
Network Function
LTE
Level ENodeBFunction
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 29

Calculate eMBMS Establishment Success Rate by using the following formula:
Figure 19 eMBMS Establishment Success Rate
— pmMbmsSessionEstabSucc
— pmMbmsSessionEstabAtt
scanner.
30 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
3 Retainability
3.1 E-RAB Retainability - Session Time Normalized Loss

Rate
This KPI measures the impact on the end user with the purpose to reflect the rate
of abnormal releases.
The Retainability Loss rate for E-RAB is given by the following equation. The PM
counters in this KPI are on cell level.
Network Function
LTE
Level EUtranCell
Calculate E-RAB Retainability - Session Time Normalized Loss Rate by using the
following formula:
Figure 20 E-RAB Retainability - Session Time Normalized Loss Rate
— pmErabRelAbnormalEnbAct
— pmErabRelAbnormalMmeAct
— pmSessionTimeUe
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 31

Note: — Since the KPI measures the impact of the network on the end user, it
also includes releases initiated by the MME. To observe the impact
of the RBS only, exclude pmErabRelAbnormalMmeAct or replace it
with pmErabRelMmeActEutra in the formula. The contribution of
several components including pmErabRelAbnormalEnbActTnFail
and pmErabRelAbnormalEnbActCdt are not part of the 3GPP
32.450 definition of the KPI. However, Ericsson thinks it is important
to include these cases even if there is no S1AP messages for these
cases (the 32.425 have its triggers built on S1AP triggers).
— To achieve the number of minutes per drop, the multiplicative

inverse of the E-RAB Retainability together with unit transformation
from seconds to minutes can be used.
— This KPI can be compensated by using flexible counters see

Compensated E-RAB Retainability - Session Time Normalized Loss
Rate.
— Cat-M1 UE devices impact on the PM counters

pmErabRelAbnormalEnbAct and pmErabRelAbnormalMmeAct
must be compensated for by using flexible counters as Cat-M1 UE
devices are not included into the PM counter pmSessionTimeUe.

E-RAB Retainability - Session Time Normalized Loss Rate.
3.2 E-RAB Retainability - Session Time Normalized Loss

Rate the Second
Network Function
LTE
Level EUtranCell
Calculate E-RAB Retainability - Session Time Normalized Loss Rate the Second
by using the following formula:
32 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
Figure 21 E-RAB Retainability - Session Time Normalized Loss Rate the Second
— pmSessionTimeUe


Compensated E-RAB Retainability - Session Time Normalized Loss
Rate the Second.
— Cat-M1 UE devices impact on the PM counter

pmErabRelAbnormalEnbAct must be compensated for by using
flexible counters as Cat-M1 UE devices are not included into the PM
counter pmSessionTimeUe.

E-RAB Retainability - Session Time Normalized Loss Rate the
Second.
3.3 Compensated E-RAB Retainability - Session Time

Normalized Loss Rate
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 33

Network Function
LTE
Level EUtranCell
Calculate Compensated E-RAB Retainability - Session Time Normalized Loss

Rate by using the following formula:

Compensated E-RAB Retainability - Session Time Normalized Loss Rate:
Figure 22 Compensated E-RAB Retainability - Session Time Normalized Loss

Rate
— pmSessionTimeUe
— pmFlexErabRelAbnormalEnbAct
— pmFlexErabRelAbnormalMmeAct
34 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability



3.4 Compensated E-RAB Retainability - Session Time

Normalized Loss Rate, the Second
Network Function
LTE
Level EUtranCell
Calculate Compensated E-RAB Retainability - Session Time Normalized Loss

Rate, the Second by using the following formula:
Figure 23 Compensated E-RAB Retainability - Session Time Normalized Loss

Rate, the Second
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 35

— pmSessionTimeUe



3.5 E-RAB Retainability - Session Time Normalized per QCI

Loss Rate
The Retainability Loss rate for an end-user service that is carried by E-RAB is
given by the following equation. The PM counters in this KPI are on cell level and
per QCI.
Network Function
LTE
Level EUtranCell
Calculate E-RAB Retainability - Session Time Normalized per QCI Loss Rate by
36 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
Figure 24 E-RAB Retainability - Session Time Normalized per QCI Loss Rate
— pmErabRelAbnormalEnbActQci
— pmErabRelAbnormalMmeActQci
— pmSessionTimeDrbQci
Note: — To achieve the number of minutes per drop, the multiplicative

— Since the KPI measures the impact of the network on the end user, it
of the RBS only, remove pmErabRelAbnormalMmeActQci in the
formula.
— The PM counters in the Equation are not included in any predefined

scanner.

pmErabRelAbnormalEnbActQci and
pmErabRelAbnormalMmeActQci must be compensated for by using
counter pmSessionTimeDrbQci see Compensated E-RAB
Retainability - Session Time Normalized Loss Rate.
of the RBS only, exclude pmErabRelAbnormalMmeActQci in the
formula.
3.6 E-RAB Retainability - Session Time Normalized per QCI

Loss Rate, the Second
The Retainability Loss rate for an end-user service that is carried by E-RAB is
given by the following equation. The PM counters in this KPI are on cell level and
per QCI.
Network Function
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 37

LTE
Level EUtranCell
Calculate E-RAB Retainability - Session Time Normalized per QCI Loss Rate, the
Second by using the following formula:
Figure 25 E-RAB Retainability - Session Time Normalized per QCI Loss Rate,
the Second
— pmSessionTimeDrbQci
Note: — To achieve the number of minutes per drop, the multiplicative

of the RBS only, remove pmErabRelAbnormalMmeActQci in the
formula.

scanner.

pmErabRelAbnormalEnbActQci and
pmErabRelAbnormalMmeActQci must be compensated for by using
counter pmSessionTimeDrbQci see Compensated E-RAB
Retainability - Session Time Normalized Loss Rate the Second.
formula.
38 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
3.7 E-RAB Retainability - Percentage Lost

This KPI measures the impact on the end user with the purpose to reflect the
percentage of established E-RABs that are lost with an abnormal release. Good
retainability is shown by a low value metric for example, < 5%.
The Retainability Loss rate for E-RAB is given by the following equations. The PM
counters in this KPI are on cell level and may be differentiated also per QCI.
Network Function
LTE
Level EUtranCell
Calculate E-RAB Retainability - Percentage Lost by using the following formula:
Figure 26 E-RAB Retainability - Percentage Lost
— pmErabRelAbnormalEnb
— pmErabRelNormalEnb
— pmErabRelMme
Note: — Several of these PM counters are not included in any predefined

scanner.

Compensated E-RAB Retainability - Percentage Lost.

E-RAB Retainability - Percentage Lost.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 39

3.8 Differentiated E-RAB Retainability - Percentage Lost

Network Function
LTE
Level EUtranCell
Calculate Differentiated E-RAB Retainability - Percentage Lost by using the

following formula:

Differentiated E-RAB Retainability - Percentage Lost:
Figure 27 Differentiated E-RAB Retainability - Percentage Lost
— pmFlexErabRelAbnormalEnb
— pmFlexErabRelNormalEnb
— pmFlexErabRelMme
40 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
Note: — Please see Detailed separation of Observability in LTE RAN for


information.

3.9 Compensated E-RAB Retainability - Percentage Lost

Network Function
LTE
Level EUtranCell
Calculate Compensated E-RAB Retainability - Percentage Lost by using the

following formula:

Compensated E-RAB Retainability - Percentage Lost:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 41

Figure 28 Compensated E-RAB Retainability - Percentage Lost
— pmErabRelMme


information.

3.10 E-RAB Retainability (eNB) - Percentage Lost

Since the KPI measures the impact of the network on the end user, it also
includes releases initiated by the MME. To observe the impact of the RBS only,
exclude pmErabRelAbnormalMmeAct or replace it with pmErabRelMmeActEutra
in the formula.
Network Function
42 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
LTE
Level EUtranCell
Calculate E-RAB Retainability (eNB) - Percentage Lost by using the following

formula:
Figure 29 E-RAB Retainability (eNB) - Percentage Lost
— pmErabRelMme
— pmErabRelMmeActEutra
of the RBS only, remove pmErabRelMmeActEutra in the formula.

Compensated E-RAB Retainability - Percentage Lost &
Compensated E-RAB Retainability(eNB) - Percentage Lost, the
Second

E-RAB Retainability - Percentage Lost & Compensated E-RAB
Retainability(eNB) - Percentage Lost, the Second.

scanner.
3.11 E-RAB Retainability (eNB) - Percentage Lost, the

Second
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 43

in the formula.
Network Function
LTE
Level EUtranCell
Calculate E-RAB Retainability (eNB) - Percentage Lost, the Second by using the
following formula:
Figure 30 E-RAB Retainability (eNB) - Percentage Lost, the Second
— pmErabRelMme
of the RBS only, remove pmErabRelMmeActEutra in the formula..

Compensated E-RAB Retainability - Percentage Lost &
Compensated E-RAB Retainability(eNB) - Percentage Lost, the
Second

E-RAB Retainability - Percentage Lost & Compensated E-RAB
Retainability(eNB) - Percentage Lost, the Second.

scanner.
44 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
3.12 Compensated E-RAB Retainability (eNB) - Percentage

Lost
in the formula.
Network Function
LTE
Level EUtranCell
Calculate Compensated E-RAB Retainability (eNB) - Percentage Lost by using


Compensated E-RAB Retainability (eNB) - Percentage Lost:
Figure 31 Compensated E-RAB Retainability (eNB) - Percentage Lost
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 45

— pmErabRelMme
— pmErabRelMmeActEutra
— pmFlexErabRelMmeActEutra


3.13 Compensated E-RAB Retainability (eNB) - Percentage

Lost, the Second
in the formula.
Network Function
LTE
Level EUtranCell
Calculate Compensated E-RAB Retainability (eNB) - Percentage Lost, the Second

by using the following formula:
46 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability

Compensated E-RAB Retainability (eNB) - Percentage Lost, the Second:
Figure 32 Compensated E-RAB Retainability (eNB) - Percentage Lost, the

Second
— pmErabRelMme


separation of Observability in LTE RAN on Dependencies and
Associated Features for more information. These flexible counters
are not included in any predefined scanner.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 47

3.14 E-RAB Retainability - Percentage Lost per QCI

percentage of established E-RAB that are lost with an abnormal release. Good
The Retainability rate for an end-user service that is carried by E-RAB is given by
the following equations. The PM counters in this KPI are on cell level and per QCI.
Network Function
LTE
Level EUtranCell
Calculate E-RAB Retainability - Percentage Lost per QCI by using the following
formula:
Figure 33 E-RAB Retainability - Percentage Lost per QCI
— pmErabRelAbnormalMmeActQci
— pmErabRelAbnormalEnbQci
— pmErabRelNormalEnbQci
— pmErabRelMmeQci
48 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
scanner.

Compensated E-RAB Retainability - Percentage Lost.

E-RAB Retainability - Percentage Lost.
formula.
3.15 E-RAB Retainability - Percentage Lost for Emergency

Calls
percentage of established E-RAB that are lost with an abnormal release. Good
The Retainability rate for Emergency Calls that is carried by E-RAB is given by
the following equation. The PM counters in this KPI are on cell level and per ARP
(hence only include the ARP range that defines Emergency).
Network Function
LTE
Level EUtranCell
Calculate E-RAB Retainability - Percentage Lost for Emergency Calls by using the
following formula:
Figure 34 E-RAB Retainability - Percentage Lost for Emergency Calls
— pmErabRelAbnormalEnbActArp
— pmErabRelAbnormalMmeActArp
— pmErabRelAbnormalEnbArp
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 49

— pmErabRelNormalEnbArp
— pmErabRelMmeArp
Note: — Several of these PM counters are not included in any predefined

scanner.
of the RBS only, exclude pmErabRelAbnormalMmeActArp in the
formula.
3.16 eMBMS Retainability - Percentage Lost

percentage of established eMBMS sessions that are lost with an abnormal
release. Good retainability is shown by a low value metric for example, < 5%.
Network Function
LTE
Level EUtranCell
Calculate eMBMS Retainability - Percentage Lost by using the following formula:
The results of the following subformulas are used in the calculation of eMBMS
Retainability - Percentage Lost:
Figure 35 eMBMS Retainability - Percentage Lost
— pmMbmsSessionEnd
— pmMbmsSessionStopAuditTimer
— pmMbmsSessionStopMme
50 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retainability
— pmMbmsSessionStopMmeNoSession
scanner.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 51

4 Integrity
4.1 Average DL UE Latency

This KPI measures the impact on the end user. The latency increases with more
instantaneously active UEs and also depends on the scheduling algorithms being
used.
DL latency for the UE is given by the following equation. The PM counters in this
KPI are on cell level and can be differentiated per QCI.
Network Function
LTE
Level EUtranCell
Calculate Average DL UE Latency by using the following formula:
Figure 36 Average DL UE Latency
— pmPdcpLatTimeDl
— pmPdcpLatPktTransDl
Note: — Cat-M1 UE devices impact in this KPI can be compensated and the
corresponding Cat-M1 only KPI is also available. See Detailed
separation of Observability in LTE RAN for related references to the
Feature Description of Category M Access.

Compensated average DL UE Latency definition. See Detailed
4.2 Compensated Average DL UE Latency

This KPI measures the impact on the end user. The latency increases with more
instantaneously active UEs and also depends on the scheduling algorithms being
used.
52 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
DL latency for the UE is given by the following equation. The PM counters in this
KPI are on cell level and can be differentiated per QCI.
Network Function
LTE
Level EUtranCell
Calculate Compensated Average DL UE Latency by using the following formula:
Figure 37 Compensated Average DL UE Latency
— pmPdcpLatTimeDl
— pmPdcpLatPktTransDl
— pmFlexPdcpLatTimeDl
— pmFlexPdcpLatPktTransDl
4.3 Average DL PDCP DRB LTE Latency per QCI

DL latency for a DRB with a given QCI value for LTE UEs. The PM counters in this
KPI are on cell level per QCI.
Network Function
LTE
Level EUtranCell
Formula Status USED
Calculate Average DL PDCP DRB LTE Latency per QCI by using the following
formula:
Figure 38 Average DL PDCP DRB LTE Latency per QCI
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 53

— pmPdcpLatTimeDlDrxNoSyncQci
— pmPdcpLatTimeDlNoDrxSyncQci
— pmPdcpLatTimeDlDrxSyncQci
— pmPdcpLatTimeDlNoDrxNoSyncQci
— pmPdcpLatPktTransDlDrxNoSyncQci
— pmPdcpLatPktTransDlNoDrxSyncQci
— pmPdcpLatPktTransDlDrxSyncQci
— pmPdcpLatPktTransDlNoDrxNoSyncQci
4.4 Average DL PDCP DRB Cat-M Latency per QCI

DL latency for a DRB with a given QCI value for Cat-M UEs. The PM counters in
this KPI are on cell level per QCI.
Network Function
LTE
Level EUtranCell
Formula Status USED
Calculate Average DL PDCP DRB Cat-M Latency per QCI by using the following
formula:
Figure 39 Average DL PDCP DRB Cat-M Latency per QCI
— pmPdcpLatTimeDlCatMDrxNoSyncQci
— pmPdcpLatTimeDlCatMNoDrxSyncQci
— pmPdcpLatTimeDlCatMDrxSyncQci
— pmPdcpLatTimeDlCatMNoDrxNoSyncQci
54 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
— pmPdcpLatPktTransDlCatMDrxNoSyncQci
— pmPdcpLatPktTransDlCatMNoDrxSyncQci
— pmPdcpLatPktTransDlCatMDrxSyncQci
— pmPdcpLatPktTransDlCatMNoDrxNoSyncQci
4.5 Average DL UE PDCP DRB Latency per QCI

DL latency for a DRB with a given QCI value is given by the following equation.
The PM counters in this KPI are on cell level per QCI.
Network Function
LTE
Level EUtranCell
Calculate Average DL UE PDCP DRB Latency per QCI by using the following
formula:
Figure 40 Average DL UE PDCP DRB Latency per QCI
— pmPdcpLatTimeDlQci
— pmPdcpLatPktTransDlQci
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 55

scanner.
— The difference in interpretation between this metric and the next

metric is shown in Figure Latency Counters.
— DL PDCP DRB latency is counted once for each PDCP active burst. A
burst is defined as time from the new data arrives to an empty PDCP
buffer until all the data from that buffer is fetched to the HARQ
buffers on the MAC layer and the PDCP buffer gets empty again.
— Flexible counters can be used to select PM counters for particular

QCIs. Filter parameters can be used for PM counters
pmFlexPdcpLatTimeDl_Filter, pmFlexPdcpLatPktTransDl_Filter. See
information. These flexible counters are not included in any
predefined scanner.

4.6 Average DL MAC DRB Latency per QCI

Average DL latency measurement per burst for a DRB with a given QCI value is
given by the following equation.
Network Function
LTE
Level EUtranCell
Formula Status DEPRECATED
Deprecated Since
L21Q1.1
Calculate Average DL MAC DRB Latency per QCI by using the following formula:
Figure 41 Average DL MAC DRB Latency per QCI
— pmMacLatTimeDlQci
56 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
— pmMacTimeDlSampQci
scanner.
— The pmMacLatTimeDlQci and pmMacTimeDlSampQci PM counters

are not available on DUL 20.
— DL MAC DRB latency is counted once for each active burst. A burst is
defined as time from the new data arrives to an empty PDCP buffer
until all the data from that buffer is successfully transmitted on the
MAC layer and both the PDCP buffer and HARQ buffers get empty.
— Different active burst definitions are used for the latency counters
pmPdcpLatTimeDlQci and pmMacLatTimeDlQci, (see Figure
Latency Counters). This means that if a new PDCP data arrives
during the transmission of the last data in the previous PDCP burst,
a new burst is considered by pmPdcpLatTimeDlQci calculations but
not by pmMacLatTimeDlQci. That is new values are added to the
counters pmPdcpLatTimeDlQci and pmPdcpLatPktTransDlQci but
not to the counters pmMacLatTimeDlQci and
pmMacTimeDlSampQci. Hence, in general, latency counters
pmPdcpLatTimeDlQci and pmMacLatTimeDlQci can count different
number of samples, especially when radio environment conditions
are bad and number of HARQ retransmissions is high.
— Flexible counters can be used to select measurements for particular

QCIs. Filter parameters can be used for PM counters
pmFlexMacLatTimeDl_Filter, pmFlexMacTimeDlSamp_Filter. See
Equation Differentiated Average DL MAC DRB Latency.

4.7 Differentiated Average DL MAC DRB Latency per QCI

Average Differentiated DL latency measurement per burst for a DRB with a given
QCI value is given by the following equation.
Network Function
LTE
Level EUtranCell
Deprecated Since
L21Q1.1
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 57

Calculate Differentiated Average DL MAC DRB Latency per QCI by using the
following formula:
Figure 42 Differentiated Average DL MAC DRB Latency per QCI
— pmFlexMacLatTimeDl
— pmFlexMacTimeDlSamp
4.8 Compensated average DL MAC DRB Latency per QCI

Average Compensated DL latency measurement per burst for a DRB with a given
QCI value is given by the following equation.
Network Function
LTE
Level EUtranCell
Deprecated Since
L21Q1.1
Calculate Compensated average DL MAC DRB Latency per QCI by using the
following formula:
Figure 43 Compensated average DL MAC DRB Latency per QCI
— pmMacLatTimeDlQci
— pmMacTimeDlSampQci
58 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
— pmFlexMacLatTimeDl
— pmFlexMacTimeDlSamp
4.9 Average DL PDCP UE Throughput

DL PDCP throughput for the UE is given by the following equation. The PM
counters in this KPI are on cell level..
Network Function
LTE
Level EUtranCell
Calculate Average DL PDCP UE Throughput by using the following formula:
Figure 44 Average DL PDCP UE Throughput
— pmUeThpTimeDl
— pmPdcpVolDlDrb
— pmPdcpVolDlDrbLastTTI
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 59

Note: — The PM counters pmPdcpVolDlDrb, pmPdcpVolDlDrbLastTTI

behaves differently in vRAN architecture see. Dependencies and
Associated Features
— For cases where pmUeThpTimeDl is small and/or where

pmPdcpVolDlDrb is similar magnitude to
pmPdcpVolDlDrbLastTtiExceptQci1 (causing a small numerator
from two large values), caution need to be used in using the metric
as the result can be ill-conditioned.

— This KPI can be compensated by using flexible counters, see

Compensated average DL PDCP UE Throughput definition. See
predefined scanner.
4.10 Compensated Average DL PDCP UE Throughput

DL PDCP throughput for the UE is given by the following equation. The PM
counters in this KPI are on cell level and can be differentiated also per QCI.
Network Function
LTE
Level EUtranCell
Calculate Compensated Average DL PDCP UE Throughput by using the following

formula:

Compensated Average DL PDCP UE Throughput:
60 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Figure 45 Compensated Average DL PDCP UE Throughput
— pmUeThpTimeDl
— pmFlexPdcpVolDlDrbLastTTI
— pmPdcpVolDlDrb
— pmPdcpVolDlDrbLastTTI
— pmFlexPdcpVolDlDrb
— pmFlexUeThpTimeDl
Note: — The PM counters pmFlexPdcpVolDlDrb,

pmFlexPdcpVolDlDrbLastTTI behaves differently in vRAN
architecture see. Dependencies and Associated Features

information.
4.11 Total DL PDCP UE Throughput

Total DL PDCP UE throughput is given by following equation which includes the
lastTTI, buffer waiting times and volume and transfer times for the last transport
blocks in all data bursts and possible HARQ retransmission delays.
Network Function
LTE
Level EUtranCell
Deprecated Since
L21Q1.1
Calculate Total DL PDCP UE Throughput by using the following formula:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 61

The results of the following subformulas are used in the calculation of Total DL
PDCP UE Throughput:
Figure 46 Total DL PDCP UE Throughput
— pmPdcpVolDlDrb
— pmMacUeThpTimeDl
Note: — pmMacUeThpTimeDl is not included in any predefined scanner
4.12 Average DL PDCP Cell Throughput

These KPIs measure the impact on the end user. DL PDCP throughput for a cell is
Network Function
LTE
Level EUtranCell
Calculate Average DL PDCP Cell Throughput by using the following formula:
Figure 47 Average DL PDCP Cell Throughput
— pmSchedActivityCellDl
— pmPdcpVolDlDrb
62 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Note: — The PM counter pmPdcpVolDlDrbQci behaves differently in vRAN

— Cell active time PM counter pmSchedActivityCellDl does not include

PDCP packet buffer waiting times (latency). The measurement
includes the transmission time for the last transport block in each
data burst.
— pmSchedActivityCellDl is not included in any predefined scanner.
— For the cells with a large amount of traffic consisting of small data
bursts. Giving an example. Web, mobile apps or VoLTE, this
throughput formula can not be a good measurement to evaluate
performance. For such traffic PDCP packet transfer time may be
dominated by buffer waiting times (latency) which is not included
into this measurement.
4.13 Average DL MAC Cell Throughput

These KPIs measure the impact on the end user. DL MAC throughput for a cell is
Network Function
LTE
Level EUtranCell
Calculate Average DL MAC Cell Throughput by using the following formula:
Figure 48 Average DL MAC Cell Throughput
— pmSchedActivityCellDl
— pmRadioThpVolDl
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 63

Note: — The volume PM counter pmRadioThpVolDl is at the MAC layer and

thus if Carrier Aggregation is active, includes both PCell traffic and
also the traffic carried as SCell traffic on behalf on another PCell.
— The PM counters in Equation are not included in any predefined

scanner.
— Cat-M1 UE devices impact is removed from the PM counters in this

KPI.
4.14 Filtered MBB UE DL PDCP Throughput

The distribution of Filtered MBB UE DL PDCP throughput is reported as an array
PM counter. This PM counter is updated based on the UE throughput in each
session as each UE session ends (or at ROP-end) and uses the underlying
measurements and nominal equation (on a per-session basis) provided in
Average DL PDCP UE Throughput on page 59.
Figure 49 Filtered DL PDCP UE Throughput Distribution
The PM counter pmUeThpDlMbbFilteredDistr filters out throughput of devices

which volume or time is below configured thresholds of the respective attributes:
— filterUeDlThpVolThr
— filterUeDlThpTimeThr
The following PM Counters are to measure the number of UEs that are not
included in pmUeThpDlMbbFilteredDistr:
pmUeExclThpDlVolte
Number of UE excluded from DL throughput calculation
due to VoLTE traffic
pmUeExclThpDlVolteMbb
Number of UE excluded from DL throughput calculation
due to VoLTE traffic but with simultaneous sizable MBB
traffic
pmUeExclThpDlSmallVol
Number of MBB UE excluded from DL throughput
calculation due to small volume
pmUeExclThpDlShortDrb
Number of MBB UE excluded from DL throughput
calculation due to short transport time
64 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Filtered MBB UE DL PDCP Throughput is calculated per percentile (10%tile，

20%tile，30%tile，40%tile, 50%tile), and include results of
pmUeThpDlMbbFilteredDistr and pmUeExclThpDlShortDrb.
Figure 50 Cumulative Distribution Function
Note: Throughput can be calculated for each percentile by using liner

interpolate within pmUeThpDlMbbFilteredDistr[Index], where Index
is the first value for which CDF[Index] > percentile.
PM counters: pmUeThpDlMbbFilteredDistr, pmUeExclThpDlShortDrb
4.15 DL PDCP UE Throughput Distribution

The distribution of UE DL PDCP throughput is reported as an array PM counter.
This PM counter is updated based on the UE throughput in each session as each
UE session ends (or at ROP-end) and uses the underlying measurements and
nominal equation (on a per-session basis) provided in Average DL PDCP UE
Throughput.
Network Function
LTE
Level EUtranCell
Calculate DL PDCP UE Throughput Distribution by using the following formula:
Figure 51 DL PDCP UE Throughput Distribution
— pmUeThp2DlDistr
Note: — The PM counter pmUeThp2DlDistr behaves differently in vRAN

4.16 DL PDCP UE Throughput Distribution for MBB UE

The mean distribution of UE DL PDCP throughput is reported as an array PM
counter. This PM counter is updated based on the UE throughput in each session
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 65

as each UE session ends (or at ROP-end) and uses the underlying measurements
and nominal equation (on a per-session basis) provided in Average DL PDCP UE
Throughput. The PM counter filters out throughput of devices which value or time
is below configured thresholds of the respective attributes. filterUeDlThpVolThr
and filterUeDlThpTimeThr.
Network Function
LTE
Level EUtranCell
Calculate DL PDCP UE Throughput Distribution for MBB UE by using the

following formula:
Figure 52 DL PDCP UE Throughput Distribution for MBB UE
— pmUeThpDlMbbFilteredDistr
4.17 Mean DL PDCP UE Throughput for MBB UE

The mean distribution of UE DL PDCP throughput is reported as an array PM
and nominal equation (on a per-session basis) provided in Average DL PDCP UE
Throughput.
The PM counter pmUeThpDlMbbFilteredDistr filters out throughput of devices

which value or time is below configured thresholds of the respective attributes.
filterUeDlThpVolThr and filterUeDlThpTimeThr.
The value for MeanBinValue[i] = 0.5 x ( Bin_Upperbound[i] +

Bin_LowerBound[i] ). Since the last bin has no static Bin_Upperbound[i] defined,
a default of 1000000 is recommended for that (and is default used by reports in
OSS, where this default can be adjusted if desired).
This value is generally different from the legacy equation in Average DL PDCP UE
Throughput as in this case each session has an equal weight in the averaging.
Both are correct, but from different viewpoints.
Network Function
LTE
Level EUtranCell
66 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Calculate Mean DL PDCP UE Throughput for MBB UE by using the following

formula:
Figure 53 Mean DL PDCP UE Throughput for MBB UE
— pmUeThpDlMbbFilteredDistr
Note: — The PM counters are not included in any predefined scanner.
— The following PM Counters are to measure throughput of UEs that

were not included in this formula.
— Number of UE excluded from DL throughput calculation due to

VoLTE traffic. pmUeExclThpDlVolte.
— Number of UE excluded from DL throughput calculation due to

VoLTE traffic but with simultaneous sizable MBB traffic.
pmUeExclThpDlVolteMbb.
— Number of MBB UE excluded from DL throughput calculation due to

small volume. pmUeExclThpDlSmallVol.
— Number of MBB UE excluded from DL throughput calculation due to

short transport time. pmUeExclThpDlShortDrb.
4.18 Mean DL PDCP UE Throughput

The distribution of UE DL PDCP throughput is reported as an array PM counter.
This PM counter is updated based on the UE throughput in each session as each
UE session ends (or at ROP-end) and uses the underlying measurements and
nominal equation (on a per-session basis) provided in Average DL PDCP UE
Throughput. From the pmUeThp2DlDistr PM counter, a mean UE PDCP
throughput (on a per-session basis) can be calculated. The value for
MeanBinValue[i] = 0.5 x ( Bin_Upperbound[i] + Bin_LowerBound[i] ). Since the
last bin has no static Bin_Upperbound[i] defined, a default of 1500000 is
recommended for that (and is default used by reports in OSS, where this default
can be adjusted if desired). This value is generally different from the legacy
equation in Average DL PDCP UE Throughput as in this case each session has an
equal weight in the averaging. Both are correct, but from different viewpoints.
Network Function
LTE
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 67

Level EUtranCell
Calculate Mean DL PDCP UE Throughput by using the following formula:
Figure 54 Mean DL PDCP UE Throughput
— pmUeThp2DlDistr
Note: — The PM counters pmUeThp2DlDistr behaves differently in vRAN

4.19 Average DL PDCP UE Throughput for Carrier

Aggregation
For traffic which is using carrier aggregation, different PM counters are provided
and can be used for that traffic based on the legacy equation in Average DL
PDCP UE Throughput
Network Function
LTE
Level EUtranCell
Calculate Average DL PDCP UE Throughput for Carrier Aggregation by using the

following formula:
Figure 55 Average DL PDCP UE Throughput for Carrier Aggregation
— pmPdcpVolDlDrbCa
— pmPdcpVolDlDrbLastTTICa
68 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
— pmUeThpTimeDlCa
Note: — The PM counters pmPdcpVolDlDrbCa, pmPdcpVolDlDrbLastTTICa

Associated Features

scanner.
— The carrier aggregation configured status per UE is only updated

once per second, so for the updating in the pmxxxCa PM counters,
there may be a slight inaccuracy for UE CA status changes within a
second. Although a UE is configured for carrier aggregation, there is
no guarantee that carrier aggregation is actually being used for that
UE in that second.
4.20 Compensated Average DL PDCP UE Throughput for

Carrier Aggregation
For traffic which is using carrier aggregation, different PM counters are provided
and can be used for that traffic based on the legacy equation in Average DL
PDCP UE Throughput.
Network Function
LTE
Level EUtranCell
Calculate Compensated Average DL PDCP UE Throughput for Carrier

Aggregation by using the following formula:

Compensated Average DL PDCP UE Throughput for Carrier Aggregation:
Figure 56 Compensated Average DL PDCP UE Throughput for Carrier

Aggregation
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 69

— pmPdcpVolDlDrbLastTTICa
— pmPdcpVolDlDrbCa
— pmUeThpTimeDlCa
— pmFlexPdcpVolDlDrbLastTTIUeCa
— pmFlexPdcpVolDlDrbCa
— pmFlexUeThpTimeDlCa
Note: — The PM counters pmPdcpVolDlDrbCa, pmPdcpVolDlDrbLastTTICa

Associated Features

scanner.

once per second, so for the updating in the pmxxxCa PM counters,
there may be a slight inaccuracy for UE CA status changes within a
UE in that second.
— These KPIs are differentiated using flexible counters. See Detailed

separation of Observability in LTE RAN for more information.
4.21 Filtered MBB UE UL Throughput

Filtered MBB UE UL Throughput is calculated per percentile (10%tile，20%tile，
30%tile，40%tile, 50%tile), and include results of
pmMacUeThpUlMbbLowVolDistr, pmUeExclThpUlShortTimeLowVol,
pmUeExclThpUlRelativeLowVol or pmMacUeThpUlMbbHighVolDistr,
pmUeExclThpUlShortTimeHighVol, pmUeExclThpUlRelativeHighVol.
Figure 57 Cumulative Distribution Function for filtered MBB UE UL Throughput

with Lower Volumes
70 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Figure 58 UeExclThpUlLow
Figure 59 Cumulative Distribution Function for filtered MBB UE UL Throughput

with Higher Volumes
Figure 60 UeExclThpUlHigh
Note: Throughput can be calculated for each percentile by using liner

interpolate within pmMacUeThpUlMbbLowVolDistr[Index] or
pmMacUeThpUlMbbHighVolDistr[Index], where Index is the first value
that satisfy CDF[Index] > percentile.
PM counters: pmMacUeThpUlMbbLowVolDistr,
pmUeExclThpUlShortTimeLowVol, pmUeExclThpUlRelativeLowVol,
pmMacUeThpUlMbbHighVolDistr, pmUeExclThpUlShortTimeHighVol,
pmUeExclThpUlRelativeHighVol
4.22 Average DL PDCP UE DRB Throughput per QCI

DL PDCP UE Throughput for both AM and UM for end-user services (defined by
QCI) is given by the following equation. The PM counters in this KPI are on cell
level and per QCI.
Network Function
LTE
Level EUtranCell
Calculate Average DL PDCP UE DRB Throughput per QCI by using the following
formula:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 71

Figure 61 Average DL PDCP UE DRB Throughput per QCI
— pmDrbThpTimeDlQci
— pmPdcpVolDlDrbLastTTIQci
— pmPdcpVolDlDrbQci
Note: — For QCI where all bearers have only ROHC traffic the KPI will show
zero. This is to filter out VoLTE which volume is measured
differently, see. pmUeExclThpDlVolte.
— The PM counters pmPdcpVolDlDrbQci, pmPdcpVolDlDrbLastTTIQci

Associated Features
— For cases where pmDrbThpTimeDlQci is small and / or where

pmPdcpVolDlDrbQci is similar magnitude to
pmPdcpVolDlDrbLastTtiQci (causing a small numerator from two
large values), caution is to be used in using the "average" metric as
the result may be ill-conditioned.

scanner.
— This KPI can be compensated by using flexible counters. Filter

parameters can be used for PM counter
pmFlexDrbThpTimeDl_Filter. See Detailed separation of
Observability in LTE RAN for more information. These flexible
counters are not included in any predefined scanner.

4.23 Average UL PDCP UE Throughput

UL PDCP throughput for the UE is given by the following equation. The PM
Network Function
LTE
72 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Level EUtranCell
Calculate Average UL PDCP UE Throughput by using the following formula:
Figure 62 Average UL PDCP UE Throughput
— pmUeThpVolUl
— pmUeThpTimeUl
Note: — For cases where pmUeThpTimeUl is small and / or where

pmUeThpVolUl is small, caution is to be used in using the metric as
the result may be ill-conditioned.
— For traffic, which uses small SDU (for example, VoLTE), the
throughput may not be a good measurement to evaluate
performance, since most of the traffic may only use a single TTI and
may be counted in the "LastTti" leaving minimal net volume if any,
on the numerator.
— If UL Carrier Aggregation is active this KPI is expected to increase.

The PM counters pmUeThpVolUl and pmUeThpTimeUl include cell
level data regardless of whether the scheduled data relates to traffic
carried on this PCell or received from an SCell. It does not show any
traffic received for another PCell, which this cell supports as an
SCell.

Compensated average UL PDCP UE Throughput. See Detailed
— Cat-M1 UE devices impact can be compensated for in this KPI by

using flexible counters, see Compensated average UL PDCP UE
Throughput definition for an example on how to use flexible
counters.

37/1553-LZA 701 6014/1 Uen H | 2021-08-25 73

4.24 Compensated Average UL PDCP UE Throughput

UL PDCP throughput for the UE is given by the following equation. The PM
Network Function
LTE
Level EUtranCell
Calculate Compensated Average UL PDCP UE Throughput by using the following

formula:
Figure 63 Compensated Average UL PDCP UE Throughput
— pmUeThpVolUl
— pmUeThpTimeUl
— pmFlexUeThpVolUl
— pmFlexUeThpTimeUl
Note: — Flexible counters can be filtered for different UE EN-DC stages. See
information.
4.25 Average UL PDCP Cell Throughput

UL PDCP throughput for a cell is given by the following equation.
Network Function
LTE
Level EUtranCell
Calculate Average UL PDCP Cell Throughput by using the following formula:
74 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Figure 64 Average UL PDCP Cell Throughput
— pmPdcpVolUlDrb
— pmSchedActivityCellUl
Note: — The PM counter pmPdcpVolUlDrb behaves differently in vRAN

— pmSchedActivityCellUl is not included in any predefined scanner.
— The volume PM counter pmPdcpVolUlDrb is measured after any

ROHC decompression (for applicable traffic streams).
— If UL Carrier Aggregation is active, the PM counters

pmPdcpVolUlDrb and pmSchedActivityCellUl include cell level data
regardless of whether the scheduled data relates to traffic carried on
this PCell or received from an SCell. It does NOT show any traffic
received for another PCell, which this cell supports as an SCell.
— Cat-M1 UE devices impact can be compensated for in this KPI by

using flexible counters on pmPdcpVolUlDrb. Filter parameters can
be used for counter pmFlexPdcpVolUlDrb. See Detailed separation
of Observability in LTE RAN for more information and an example
on how to use flexible counters. Cat-M1 UEs are not included into
the counter pmSchedActivityCellUl. These flexible counters are not
included in any predefined scanner.
4.26 Average UL MAC Cell Throughput

UL MAC throughput for a cell is given by the following equation.
Network Function
LTE
Level EUtranCell
Calculate Average UL MAC Cell Throughput by using the following formula:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 75

Figure 65 Average UL MAC Cell Throughput
— pmSchedActivityCellUl
— pmRadioThpVolUl
scanner.
— The volume PM counter pmRadioThpVolUl is at the MAC layer and

thus if UL Carrier Aggregation is active, includes both PCell traffic
and also the traffic carried as SCell traffic on behalf on another
PCell.
— Cat-M1 UE devices impact is removed from the PM counters in this

KPI.
4.27 Normalized Average UL MAC Cell Throughput

Considering Successful PUSCH subframe Only
Normalized average UL MAC throughput for a cell, based on successfully received
SDU on PUSCH. Given by the following equation.
Network Function
LTE
Level EUtranCell
Formula Status USED
Calculate Normalized Average UL MAC Cell Throughput Considering Successful

PUSCH subframe Only by using the following formula:
Figure 66 Normalized Average UL MAC Cell Throughput Considering Successful

PUSCH subframe Only
76 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
— pmMacVolUl
— pmPuschSchedActivity
Note: — The PM counters in the equation are not included in any predefined
scanner.
— The pmMacVolUl counter is at the MAC layer. If Uplink Carrier

Aggregation is active, the counters include volume transmitted on
this cell regardless of whether it is used as PCell or SCell and the
same for time in pmPuschSchedActivity.
— This KPI is only showing non-Category M traffic.This KPI is

recommended when using ESS.
4.28 Normalized Average UL MAC Cell Throughput

Considering Successful PUSCH Slot Only
This KPI is deprecated and replaced by "Normalized Average UL MAC Cell
Throughput Considering Successful PUSCH subframe Only".
Network Function
LTE
Level EUtranCell
Deprecated Since
L20Q4.1
Calculate Normalized Average UL MAC Cell Throughput Considering Successful

PUSCH Slot Only by using the following formula:
Figure 67 Normalized Average UL MAC Cell Throughput Considering Successful

PUSCH Slot Only
— pmMacVolUl
— pmPuschSchedActivity
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 77

Note: — The PM counters in the equation are not included in any predefined
scanner.
— The pmMacVolUl counter is at the MAC layer. If Uplink Carrier

Aggregation is active, the counters include volume transmitted on
this cell regardless of whether it is used as PCell or SCell and the
same for time in pmPuschSchedActivity.
— This KPI is only used for non-Category M traffic.This KPI is

recommended when using ESS.
4.29 Average UL PDCP UE Throughput for Carrier

Aggregation
For traffic which is in a CA_Configured state, different PM counters are provided
and can be used for that traffic based on the legacy equation in Average UL
PDCP UE Throughput.
Network Function
LTE
Level EUtranCell
Calculate Average UL PDCP UE Throughput for Carrier Aggregation by using the

following formula:
Figure 68 Average UL PDCP UE Throughput for Carrier Aggregation
— pmUeThpVolUlCa
— pmUeThpTimeUlCa
Note: — This KPI can be compensated by using flexible counters, see

Equation Compensated average CA UL PDCP UE Throughput. See
information.
78 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
4.30 Compensated Average UL PDCP UE Throughput for

Carrier Aggregation
For traffic which is in a CA_Configured state, different PM counters are provided
and can be used for that traffic based on the legacy equation in Average UL
PDCP UE Throughput.
Network Function
LTE
Level EUtranCell
Calculate Compensated Average UL PDCP UE Throughput for Carrier

Aggregation by using the following formula:
Figure 69 Compensated Average UL PDCP UE Throughput for Carrier

Aggregation
— pmUeThpVolUlCa
— pmUeThpTimeUlCa
— pmFlexUeThpVolUlCa
— pmFlexUeThpTimeUlCa
scanner.

once per second, so for the updating in the pmxxxCa counters, there
may be a slight inaccuracy for UE CA status changes within a
UE in that second.

information.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 79

4.31 Average MAC UE UL THROUGHPUT

Uplink user throughput for traffic restricted by the air interface. The KPI fulfils
"Scheduled IP Throughput" in 3GPP TS 36.314.
Network Function
LTE
Level EUtranCell
Calculate Average MAC UE UL THROUGHPUT by using the following formula:
Figure 70 Average MAC UE UL THROUGHPUT
— pmMacVolUlResUe
— pmMacTimeUlResUe
4.32 UL PDCP UE Throughput Distribution

The distribution of UE uplink PDCP throughput is reported as an array PM
and nominal equation (on a per-session basis) provided in Average UL PDCP UE
Throughput.
Network Function
LTE
Level EUtranCell
Calculate UL PDCP UE Throughput Distribution by using the following formula:
Figure 71 UL PDCP UE Throughput Distribution
80 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
— pmUeThp2UlDistr
4.33 Mean UL PDCP UE Throughput

The distribution of UE uplink PDCP throughput is reported as an array PM
and nominal equation (on a per-session basis) provided in Average UL PDCP UE
Throughput.
From the pmUeThp2UlDistr PM counter, a mean UE PDCP throughput (on a per-

session basis) can be calculated.
The value for MeanBinValue[i] = 0.5 x ( Bin_Upperbound[i] +

Bin_LowerBound[i] ). Since the last bin has no static Bin_Upperbound[i] defined,
a default of 200000 is recommended for that and is default used by reports in
OSS, where this default can be adjusted if desired.
Network Function
LTE
Level EUtranCell
Calculate Mean UL PDCP UE Throughput by using the following formula:
Figure 72 Mean UL PDCP UE Throughput
— pmUeThp2UlDistr
Note: — This value is in general different from the legacy equation in

Average UL PDCP UE Throughput as in this case each session has
an equal weight in the averaging. Both are correct, but from
different viewpoints.
4.34 Average UL UE Throughput per LCG

This KPI measures the impact on the end user. For traffic which use small SDU
(for example, VoLTE), the throughput may not be a good measurement to
evaluate performance, since most of the traffic may only use a single TTI and
may be excluded from the PM counters.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 81

For the UL, instead of measuring per QCI the measurements are Logical Channel
Group (LCG) level. For and the connection between LCG and QCI, see Quality of
Service. UL throughput for the UE per LCG is given by the following equation. The
PM counters in this KPI are on cell level and per LCG.
Network Function
LTE
Level EUtranCell
Calculate Average UL UE Throughput per LCG by using the following formula:
Figure 73 Average UL UE Throughput per LCG
— pmLcgThpVolUlLcg
— pmLcgThpTimeUlLcg
Note: — For cases where pmLcgThpTimeUlLcg is small and / or where

pmLcgThpVolUlLcg is small, caution is to be used in using the metric
as the result may be ill-conditioned.
— For traffic, which uses small SDU (for example, VoLTE), the
throughput may not be a good measurement to evaluate
performance, since most of the traffic may only use a single TTI and
may be counted in the "LastTti" leaving minimal net volume if any,
on the numerator.


Equation Compensated average UL Throughput per LCG. See
predefined scanner.
4.35 Compensated average UL Throughput per LCG

This KPI measures the impact on the end user. For traffic which use small SDU
(for example, VoLTE), the throughput may not be a good measurement to
82 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
evaluate performance, since most of the traffic may only use a single TTI and
may be excluded from the PM counters.
For the UL, instead of measuring per QCI the measurements are Logical Channel
Group (LCG) level. For and the connection between LCG and QCI, see Quality of
Service. UL throughput for the UE per LCG is given by the following equation. The
PM counters in this KPI are on cell level and per LCG.
Network Function
LTE
Level EUtranCell
Calculate Compensated average UL Throughput per LCG by using the following

formula:
Figure 74 Compensated average UL Throughput per LCG
— pmLcgThpVolUlLcg
— pmLcgThpTimeUlLcg
— pmFlexLcgThpVolUl
— pmFlexLcgThpTimeUl
4.36 Average DL Packet Error Loss Rate

DL packet error loss rate for the UE, from a cell perspective, is given by the
following equation. The PM counters in this KPI are on several MO classes.
Network Function
LTE
Level EUtranCell
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 83

Calculate Average DL Packet Error Loss Rate by using the following formula:
The results of the following subformulas are used in the calculation of Average
DL Packet Error Loss Rate:
Figure 75 Average DL Packet Error Loss Rate
— pmPdcpPktDiscDlHo
— pmPdcpPktReceivedDl
— pmPdcpPktFwdDl
— pmPdcpPktDiscDlPelr
— pmPdcpPktDiscDlPelrUu
— pmPdcpPktDiscDlEth
— pmPdcpPktDiscDlNoUeCtxt
— pmMbmsPktRecSynchPayload
— pmMbmsPktRecSynchCtrl
— pmPdcpPktDiscDlNoMbmsCtxt
84 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Note: — The term " A " represents packets discarded (pmPdcpPktDiscDlEth,

pmPdcpPktDiscDlNoUeCtxt) which cannot be mapped to any
specific cell. Therefore an estimate is used, consisting of weighting
the total number of these discarded packets with the fraction of
packets that went to that cell versus all packets on all cells of the
RBS. Term "C" similarly removes a fraction of the discards
apportioned to eMBMS. If eMBMS is inactive, C wil be equal to 1
and pmPdcpPktDiscDlEth will not be reduced.

scanner.
— This KPI can be compensated by using flexible counters. The filter

parameters can be used for flexible counters,
pmFlexPdcpPktDiscDlPelrUu_Filter, pmFlexPdcpPktDiscDlHo_Filter,
pmFlexPdcpPktReceivedDl_Filter, pmFlexPdcpPktFwdDl_Filter. See
information and an example on how to use flexible counters. These

4.37 Average DL Packet Error Loss Rate per QCI

DL packet error loss rate for the UE is given by the following equation. The PM
counters in this KPI are on several MO classes. It is important to aggregate them
on a common basis (for example, cell level) to obtain a meaningful metric.
Network Function
LTE
Level EUtranCell
Calculate Average DL Packet Error Loss Rate per QCI by using the following
formula:
DL Packet Error Loss Rate per QCI:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 85

Figure 76 Average DL Packet Error Loss Rate per QCI
— pmPdcpPktTransDlQci
— pmPdcpPktDiscDlPelrQci
— pmPdcpPktDiscDlPelrUuQci
— pmPdcpPktDiscDlHoQci
— pmPdcpPktDiscDlEth
— pmPdcpPktDiscDlNoUeCtxt
— pmPdcpPktReceivedDlQci
— pmPdcpPktReceivedDl
— pmMbmsPktRecSynchCtrl
— pmPdcpPktDiscDlNoMbmsCtxt
86 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Note: — The term " B " represents packets discarded (pmPdcpPktDiscDlEth,

pmPdcpPktDiscDlNoUeCtxt) which cannot be mapped to any
specific cell. Therefore an estimate is used, consisting of weighting
the total number of these discarded packets with the fraction of the
specific QCI packets that went to that cell versus all packets on all
cells of the RBS. Term " C " similarly removes a fraction of the
discards apportioned to eMBMS. If eMBMS is inactive, C wil be
equal to 1 and pmPdcpPktDiscDlEth will not be reduced.

scanner.
— PM counter pmPdcpPktDiscDlPelrUuQci is estimated from overall

average HARQ failure statistics, so its value may be inappropriately
assigned to specific QCI
— This KPI can be compensated by using flexible counters. Filter

parameters can be used for PM counters
pmFlexPdcpPktDiscDlPelr_Filter,
pmFlexPdcpPktDiscDlPelrUu_Filter, pmFlexPdcpPktDiscDlHo_Filter,
pmFlexPdcpPktReceivedDl_Filter. See Detailed separation of
Observability in LTE RAN for more information and an example on
how to use flexible counters. These flexible counters are not
included in any predefined scanner.

4.38 Average eMBMS Packet Loss Rate

The eMBMS packet loss rate for the UE is given by the following equation. The
PM counters in this KPI are on the MbmsService MO level.
Network Function
LTE
Level ENodeBFunction
Calculate Average eMBMS Packet Loss Rate by using the following formula:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 87

eMBMS Packet Loss Rate:
Figure 77 Average eMBMS Packet Loss Rate
— pmMbmsPktDiscSynchPayloadMuted
— pmMbmsPktDiscSynchPayloadTooMuch
scanner.
4.39 Average UL Packet Loss Rate

Corresponding other KPI recommended for vRAN architecture exist, see.

Dependencies and Associated Features.
UL packet loss rate for the UE is given by the following equation. The PM
Network Function
LTE
Level EUtranCell
Calculate Average UL Packet Loss Rate by using the following formula:
Figure 78 Average UL Packet Loss Rate
— pmPdcpPktLostUl
— pmPdcpPktReceivedUl
88 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Note: — This KPI can be compensated by using flexible counters. Filter

parameters can be used for PM counters
pmFlexPdcpPktLostUl_Filter, pmFlexPdcpPktReceivedUl. See
information and an example on how to use flexible counters. These

4.40 Average UL Packet Loss Rate per QCI

Corresponding other KPI recommended for vRAN architecture exist, see.

Dependencies and Associated Features.
UL packet loss rate for the UE is given by the following equation. The PM
counters in this KPI are on cell level and per QCI.
Network Function
LTE
Level EUtranCell
Calculate Average UL Packet Loss Rate per QCI by using the following formula:
Figure 79 Average UL Packet Loss Rate per QCI
— pmPdcpPktLostUlQci
— pmPdcpPktReceivedUlQci
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 89

scanner.

4.41 VoIP Cell Integrity

This KPI measures the integrity impact for VoIP in the RAN.
An E-RAB is defined as VoIP based on settings of the QciProfilePredefined /

QciProfileOperatorDefined MO attribute ServiceType.
Simulations have shown that the UL is usually the limiting factor. VoIP integrity is
a measurement that shows how many % of the VoIP users that have 99% of their
UL packets within the packet delay budget. The satisfaction rate for the UE is
given by the following equation. The PM counters in this KPI are on cell level.
Network Function
LTE
Level EUtranCell
Calculate VoIP Cell Integrity by using the following formula:
Figure 80 VoIP Cell Integrity
— pmVoipQualityRbUlOk
— pmVoipQualityRbUlNok
90 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
Note: — The PM counters pmVoipQualityRbUlOk, pmVoipQualityRbUlNok

Associated Features
— This KPI can be compensated by using flexible counters. See

information and an example on how to use flexible counters.

scanner.
4.42 Compensated VoIP Cell integrity

This KPI measures the integrity impact for VoIP in the RAN.
An E-RAB is defined as VoIP based on settings of the QciProfilePredefined /

QciProfileOperatorDefined MO attribute ServiceType.
Simulations have shown that the UL is usually the limiting factor. VoIP integrity is
a measurement that shows how many % of the VoIP users that have 99% of their
UL packets within the packet delay budget. The satisfaction rate for the UE is
given by the following equation. The PM counters in this KPI are on cell level.
Network Function
LTE
Level EUtranCell
Calculate Compensated VoIP Cell integrity by using the following formula:

Compensated VoIP Cell integrity:
Figure 81 Compensated VoIP Cell integrity
— pmVoipQualityRbUlOk
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 91

— pmVoipQualityRbUlNok
— pmFlexVoipQualityRbUlOk
— pmFlexVoipQualityRbUlNok
Note: — The PM counters pmVoipQualityRbUlOk, pmVoipQualityRbUlNok

Associated Features

scanner.
4.43 GTP-U Downlink Packet Loss Rate

This KPI measures the network performance. PM counters in this KPI are on QCI
level.
GTP-U Downlink Packet Loss Rate shows percentage of lost GTP-U user packets
in downlink per QCI.
Network Function
LTE
Level TermPointToSGW
Calculate GTP-U Downlink Packet Loss Rate by using the following formula:
Figure 82 GTP-U Downlink Packet Loss Rate
— pmS1uPktLostDlQci
— pmS1uPktRecQci
Note: — For cases where sequence number in received GTP-U packets on S1-
U is not present the metric as the result may be ill-conditioned.
4.44 GTP-U Downlink Packet Out of Order Rate

This KPI measures the network performance. PM counters in this KPI are on QCI
level.
92 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Integrity
GTP-U Downlink Packet Out Of Order Rate shows percentage of GTP-U user
packets received out of order in downlink per QCI.
Network Function
LTE
Level TermPointToSGW
Calculate GTP-U Downlink Packet Out of Order Rate by using the following
formula:
Figure 83 GTP-U Downlink Packet Out of Order Rate
— pmS1uPktRecOooQci
— pmS1uPktRecQci
Note: — For cases where sequence number in received GTP-U packets on S1-
U is not present the metric as the result may be ill-conditioned.
4.45 DL Scheduled Entities per TTI

The following KPIs are to measure scheduled entities per TTI in Downlink. The
PM counters in this KPI are on cell level.
Network Function
LTE
Level EUtranCell
Calculate DL Scheduled Entities per TTI by using the following formula:
Figure 84 DL Scheduled Entities per TTI
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 93

— pmSchedActivityUeDl
— pmSchedTtiCellDl
Note: — These PM Counters are not included in any predefined scanner.
4.46 UL Scheduled Entities per TTI

The following KPIs are to measure scheduled entities per TTI in Uplink. The PM
Network Function
LTE
Level EUtranCell
Calculate UL Scheduled Entities per TTI by using the following formula:
Figure 85 UL Scheduled Entities per TTI
— pmSchedActivityUeUl
— pmSchedTtiCellUl
Note: — These PM Counters are not included in any predefined scanner.
94 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility
5 Mobility
5.1 Mobility Success Rate

This KPI measures system performance. Mobility Success rate includes both
preparation of target cell resources and move from the source cell to the target
cell, as given in the following equation. PM counters in this KPI are on the MO
EUtranCellRelation or UtranCellRelation or GeranCellRelation or
Cdma2001xRttCellRelation level and must thus be aggregated over the various
cell pairings for each type of handover.
Network Function
LTE
Level EUtranCellRelation
Calculate Mobility Success Rate by using the following formula:
The results of the following subformulas are used in the calculation of Mobility
Success Rate:
Figure 86 Mobility Success Rate
— pmHoPrepSuccLteIntraF
— pmHoPrepSuccLteInterF
— pmHoPrepSucc
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 95

— pmHoPrepSucc1xRttSrvcc
— pmHoPrepSuccNonMob
— pmHoPrepAttLteIntraF
— pmHoPrepAttLteInterF
— pmHoPrepAtt
— pmHoPrepAtt1xRttSrvcc
— pmHoPrepAttNonMob
— pmHoPrepRejInUlThres
— pmHoExeSuccLteIntraF
— pmHoExeSuccLteInterF
— pmHoExeSucc
— pmHoExeSuccNonMob
— pmHoExeAttLteIntraF
— pmHoExeAttLteInterF
— pmHoExeAtt
— pmHoExeAttNonMob
96 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility
Note: — If handover for a UE fails, but is followed by a successful

reestablishment, the handover is still treated as unsuccessful by this
KPI.
— Some of the features (for example, Load balancing and eMBMS

related features) can trigger handover without any mobility. So PM
counters with "NonMob" suffix (pmHoPrepSuccNonMob,
pmHoPrepAttNonMob, pmHoExeSuccNonMob, &
pmHoExeAttNonMob) remove handovers that are triggered with
those non-mobility factors rather than mobility.
— The PM counters pmHoPrepSucc, pmHoPrepAtt1xRttSrvcc,

pmHoPrepAtt, pmHoExeSucc, pmHoPrepSucc1xRttSrvcc &
pmHoExeAtt refer to handovers to non-LTE systems and are NOT
parents of the corresponding LTE handover counters. PM counters in
the Equations exist in class UtranCellRelation for WCDMA and
TDScdma handovers... in class GeranCellRelation for GERAN SRVCC
handovers (see the SRVCC Handover to GERAN feature) and in
class Cdma20001xRttCellRelation for SRVCC handover to CDMA
1xRtt (see the SRVCC Handover to CDMA 1X feature).

scanner.
5.2 Cell Mobility Success Rate

This KPI measures system performance.
Rarely used cell-relations are not populated within MO EUtranCellRelation if

attribute AnrFunction.nrHONeededToAddCellRelation > 0 and thus those new
relations would NOT appear in the preceding Mobility equation until its handover
attempt usage exceeded the new attribute value. PM counters based on
EUtranCellFDD/EUtranCellTDD are cell-based. These new PM counters are
pegged whether the CellRelation has been created or not. If
AnrFunction.nrHONeededToAddCellRelation = 0 then the new PM counters and
the cell-aggregated (CellRelation) counters are equivalent. These PM counters
permit the following Cell Mobility KPI metric.
Network Function
LTE
Level EUtranCell
Calculate Cell Mobility Success Rate by using the following formula:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 97

The results of the following subformulas are used in the calculation of Cell
Mobility Success Rate:
Figure 87 Cell Mobility Success Rate
— pmCellHoPrepSuccLteIntraF
— pmCellHoPrepSuccLteInterF
— pmHoPrepSucc
— pmHoPrepSuccNonMob
— pmCellHoPrepAttLteIntraF
— pmCellHoPrepAttLteInterF
— pmHoPrepAtt
— pmHoPrepAttNonMob
— pmCellHoExeSuccLteIntraF
— pmCellHoExeSuccLteInterF
— pmHoExeSucc
98 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility
— pmHoExeSuccNonMob
— pmCellHoExeAttLteIntraF
— pmCellHoExeAttLteInterF
— pmHoExeAtt
— pmHoExeAttNonMob

KPI.
— Some of the features (for example, Load balancing and eMBMS

related features) can trigger handover without any mobility. So PM
counters with "NonMob" suffix (pmHoPrepSuccNonMob,
pmHoPrepAttNonMob, pmHoExeSuccNonMob, &
pmHoExeAttNonMob) remove handovers that are triggered with
those non-mobility factors rather than mobility.

the Equations exist in class UtranCellRelation for WCDMA and

scanner.
— The Lb, GERAN, UTRAN, and 1xRTT PM counters must be

aggregated to the cell level for these Cell Mobility equations.
— This KPI can be compensated with use of flexible counters. See

information and an example of how to use flexible counters. These

scanner.
5.3 Flexible Intra-Frequency Cell Mobility Success Rate

This KPI measures system performance. This KPI is a flexible mobility
measurement, when used with Cat-M1 filter it is dedicated for Cat-M1 devices.
The pmCounters are on cell level. These PM counters permit the following Cell
Mobility KPI metric.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 99

Network Function
LTE
Level EUtranCell
Calculate Flexible Intra-Frequency Cell Mobility Success Rate by using the

following formula:
The results of the following subformulas are used in the calculation of Flexible
Intra-Frequency Cell Mobility Success Rate:
Figure 88 Flexible Intra-Frequency Cell Mobility Success Rate
— pmFlexCellHoPrepSuccLteIntraF
— pmFlexCellHoPrepAttLteIntraF
— pmFlexCellHoExeSuccLteIntraF
— pmFlexCellHoExeAttLteIntraF
100 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility

KPI.
— This KPI is differentiated with use of flexible counters. See Detailed

separation of Observability in LTE RAN for related for more
information. For example how to use flexible counters see Detailed
separation of Observability in LTE RAN. These flexible counters are
not included in any predefined scanner.

5.4 Handover Success Rate

This KPI measures system performance. The Handover Success rate includes
both mobility triggered handovers and handovers that are not triggered with
mobility (in example for load balancing). The KPI covers both preparation of
target cell resources and move from the source cell to the target cell, as given in
the following equation. PM counters in this KPI are on the MO
Network Function
LTE
Calculate Handover Success Rate by using the following formula:
The results of the following subformulas are used in the calculation of Handover
Success Rate:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 101

Figure 89 Handover Success Rate
— pmHoPrepSuccLteIntraF
— pmHoPrepSuccLteInterF
— pmHoPrepSucc
— pmHoPrepAttLteIntraF
— pmHoPrepAttLteInterF
— pmHoPrepAtt
— pmHoExeSucc
— pmHoExeAtt
102 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility

KPI.

the Equation exist in class UtranCellRelation for WCDMA and
— Several of flexible counters are available to compensate this KPI.

See Detailed separation of Observability in LTE RAN for more
information.

scanner.
5.5 Handover Execution Success Rate

The Handover Execution Success rate includes both mobility triggered handovers
and handovers that are not triggered with mobility (in example for load
balancing). The KPI covers only the move from the source cell to the target cell,
as given in the following equation. PM counters in this KPI are on the MO
Cdma2001xRttCellRelatio level and must thus be aggregated over the various
Network Function
LTE
Calculate Handover Execution Success Rate by using the following formula:
The results of the following subformulas are used in the calculation of Handover
Execution Success Rate:
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 103

Figure 90 Handover Execution Success Rate
— pmHoExeSucc
— pmHoExeAtt
Note: — If handover for a UE fails, but is followed by a successful re-

establishment, the handover is still treated as unsuccessful by this
KPI.
— Some features (for example, Load balancing and eMBMS features)

can trigger handover without any mobility.
— The PM counters pmHoExeSucc and pmHoExeAtt refer to

handovers to non-LTE systems and are NOT parents of the
corresponding LTE handover counters. PM counters in the Equations
exist in class UtranCellRelation for WCDMA and TDScdma
handovers... in class GeranCellRelation for GERAN SRVCC

scanner.
5.6 Cell Handover Success Rate

The Handover Success rate includes both mobility triggered handovers and
handovers that are not triggered with mobility (in example for load balancing).
The KPI covers both preparation of target cell resources and move from the
source cell to the target cell, as given in the following equation. PM counters in
this KPI are on the MO EUtranCellRelation or UtranCellRelation or
GeranCellRelation or Cdma2001xRttCellRelation level and must thus be
aggregated over the various cell pairings for each type of handover.
104 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility

relations would NOT appear in the preceding Mobility equation until its handover
attempt usage exceeded the new attribute value. PM counters based on
EUtranCellFDD/EUtranCellTDD are cell-based. These new PM counters are
pegged whether the CellRelation has been created or not. If
AnrFunction.nrHONeededToAddCellRelation = 0 then the new PM counters and
the cell-aggregated (CellRelation) counters are equivalent. These PM counters
permit the Cell Handover KPI metric.
Network Function
LTE
Level EUtranCell
Calculate Cell Handover Success Rate by using the following formula:
Handover Success Rate:
Figure 91 Cell Handover Success Rate
— pmCellHoPrepSuccLteIntraF
— pmCellHoPrepSuccLteInterF
— pmHoPrepSucc
— pmCellHoPrepAttLteIntraF
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 105

— pmCellHoPrepAttLteInterF
— pmHoPrepAtt
— pmHoPrepRejOutUlThres
— pmHoExeSucc
— pmHoExeAtt

KPI.

aggregated to the cell level for these Cell Handover equations.
— Several of flexible counters are available to compensate this KPI.

See Detailed separation of Observability in LTE RAN for more
information.

scanner.
5.7 Cell Handover Execution Success Rate

The Handover Execution Success rate includes both mobility triggered handovers
and handovers that are not triggered with mobility (in example for load
balancing). The KPI covers only the move from the source cell to the target cell,
as given in the following equation. PM counters in this KPI are on the MO

relations would NOT appear in the preceding Handover Execution equation until
its handover attempt usage exceeded the new value of the attribute. A set of PM
counters based on EUtranCellFDD/EUtranCellTDD is cell-based. These new PM
counters are pegged whether the CellRelation has been created or not. If
106 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility
AnrFunction.nrHONeededToAddCellRelation = 0 then the PM counters and the

cell-aggregated (CellRelation) counters be equivalent. These PM counters permit
the following Cell Handover Execution KPI metric.
Network Function
LTE
Level EUtranCell
Calculate Cell Handover Execution Success Rate by using the following formula:
Handover Execution Success Rate:
Figure 92 Cell Handover Execution Success Rate
— pmHoExeSucc
— pmHoExeAtt

KPI.

aggregated to the cell level for these Cell Handover Execution
equations.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 107

5.8 Differentiated Cell Handover Success Rate

This KPI measures system performance.This KPI measures Cell Handover
Success Rate. The PM counters in this KPI are on cell level.
Network Function
LTE
Level EUtranCell
Calculate Differentiated Cell Handover Success Rate by using the following

formula:

Differentiated Cell Handover Success Rate:
Figure 93 Differentiated Cell Handover Success Rate
— pmFlexCellHoExeAttGeran
— pmFlexCellHoExeAttLteInterF
— pmFlexCellHoExeAttLteIntraF
— pmFlexCellHoExeSuccGeran
— pmFlexCellHoExeSuccLteInterF
— pmFlexCellHoExeSuccLteIntraF
108 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility
— pmFlexCellHoExeSuccNonMobIrat
— pmFlexCellHoExeSuccNonMobLte
— pmFlexCellHoPrepAtt1xRttSrvcc
— pmFlexCellHoPrepAttGeran
— pmFlexCellHoPrepAttLteInterF
— pmFlexCellHoPrepAttLteIntraF
— pmFlexCellHoPrepAttNonMobIrat
— pmFlexCellHoPrepAttNonMobLte
— pmFlexCellHoPrepSucc1xRttSrvcc
— pmFlexCellHoPrepSuccGeran
— pmFlexCellHoPrepSuccLteInterF
— pmFlexCellHoPrepSuccLteIntraF
— pmFlexCellHoPrepSuccNonMobIrat
— pmFlexCellHoPrepSuccNonMobLte
— pmFlexCellHoExeAttNonMobLte
— pmFlexCellHoExeAttNonMobIrat
— pmFlexCellHoPrepSuccUtran
— pmFlexCellHoPrepAttUtran
— pmFlexCellHoExeSuccUtran
— pmFlexCellHoExeAttUtran
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 109


KPI.

aggregated to the cell level for these Cell Handover Execution
equations.
— The KPIs can be differentiated with flexible counters.
— Filter parameters can be used for flexible counters. For more

information, see the 'Detailed separation of Observability in LTE
RAN' chapter of the 'Key Performance Indicators' CPI
document.These flexible counters are not included in any predefined
scanner.
— Flexible counters can be filtered for different UE EN-DC stages. For

more information, see the 'Detailed separation of Observability in
LTE RAN' chapter of the 'Key Performance Indicators' CPI
document.
5.9 UTRAN SRVCC Success Rate

This KPI measures system performance, when SRVCC occurs to an UTRAN
system
Network Function
LTE
Level UtranCellRelation
Calculate UTRAN SRVCC Success Rate by using the following formula:
Figure 94 UTRAN SRVCC Success Rate
— pmHoPrepSuccSrvcc
— pmHoPrepAttSrvcc
— pmHoExeSuccSrvcc
110 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Mobility
— pmHoExeAttSrvcc
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 111

6 Availability
6.1 Cell Availability

This KPI measures system performance. Since the KPI is measured by the RBS, it
does not include time when the RBS is down, that is, node restart time is
excluded.
The length of time in seconds that a cell is available for service is defined as cell
availability. Cell availability for a cluster of 'CELL' number of cells during 'ROP'
reporting periods can be calculated using the following formula.
Network Function
LTE
Level EUtranCell
Calculate Cell Availability by using the following formula:
Figure 95 Cell Availability
— pmCellDowntimeAuto
— pmCellDowntimeMan
112 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Availability
Note: — The manual blocking time of a cell is included in this KPI to show
the overall availability of the cell. To remove the manual
intervention impact on cell availability, remove the PM counter
pmCellDowntimeMan from the numerator and subtract the value of
the PM counter pmCellDowntimeMan from the denominator.
— If the files with the PM counters are missing, the time that those
files represent in "CELL x ROP x 900" shall be excluded from Cell
Availability result.
— A cell restart because of the Advanced Cell Supervision feature

increments the PM counter pmCellDowntimeMan.
— The default setting for the delay timer is 0, which does not affect the
cell restart time. If the delay timer is set to be larger than 0, cell
restart/unlock time duration is extended as a consequence of the
number of TX updates taken place during the lock/unlock
procedure.
— Downtime because of Cell Sleep mode is not impacting the user

perceived availability and therefore pmCellSleepTime is not
included in this KPI.
— If Clock is adjusted by more than 300 seconds in order to keep up

with the time references, and the PM counters
pmCellDowntimeAuto/pmCellDowntimeMan step above 1200,
those PM counters are set to blank value in current ROP.
37/1553-LZA 701 6014/1 Uen H | 2021-08-25 113

L21Q3KPI

Uploaded by

Copyright:

Available Formats

L21Q3KPI

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

L21Q3KPI

Uploaded by

Copyright:

Available Formats

Key Performance Indicators

37/1553-LZA 701 6014/1 Uen H

© Ericsson AB 2021. All rights reserved. No part of this document may be

37/1553-LZA 701 6014/1 Uen H | 2021-08-25

37/1553-LZA 701 6014/1 Uen H | 2021-08-25

3.4 Compensated E-RAB Retainability - Session Time Normalized

37/1553-LZA 701 6014/1 Uen H | 2021-08-25

4.20 Compensated Average DL PDCP UE Throughput for Carrier

37/1553-LZA 701 6014/1 Uen H | 2021-08-25

5.7 Cell Handover Execution Success Rate 106

37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Monitoring the performance of a RAN through KPIs enables the following

— Detecting performance issues in the network to preserve service quality.

— Providing radio network planners with detailed information for

— Troubleshooting on cell clusters of interest.

Performance Management (PM) counters collect performance statistics from

1.1 End-User Performance in LTE

Application UE RBS TN SGW TN PGW Server

PGW Packet Data Network Gateway

SGW Serving Gateway

37/1553-LZA 701 6014/1 Uen H | 2021-08-25 1

Figure 1 LTE Packet Delivery Chain

1.2 Observability in Ericsson LTE

Key Performance Indicator (KPI) level End-user

Performance Indicator (PI) level

Figure 2 Top-Down Approach in LTE RAN Observability

Table 1 Observability Levels and Purposes

2 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Observability Level Description of Use

1.3 LTE RAN Performance Observability Model

Table 2 Ericsson Observability Model

37/1553-LZA 701 6014/1 Uen H | 2021-08-25 3

QoS Categories Description

1.4 Aggregation of KPIs

— Over a cluster of cells

— Over one RBS

— Over multiple RBSs

— Over a complete geographic area

4 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

— Over time periods (hour, days, week)

Aggregation needs to be considered for maximum calculations in case of some

2. Perform the PI equation.

In certain situations, an operator needs to replace the individual pmXyz PM

37/1553-LZA 701 6014/1 Uen H | 2021-08-25 5

— Use of PI(Cell&ROP) can highlight the impact of individual metrics as each

— In cases where the denominator pmRst(Cell&ROP) is a fixed number in the

— In case of throughput metrics, SUM(all_Volume) / SUM(all_Time), can

— Aggregation of quantities based on time (for example, pmUeThpTimeDl and

— Same aggregation methodology needs to be used comparing results to avoid

— SUM function needs to be avoided, if the aggregated quantities are

Aggregation is indicated in the formulas when in use.

6 37/1553-LZA 701 6014/1 Uen H | 2021-08-25

Retention Priority (ARP), then it is possible to determine the observability of the

Web sessions If seven is defined as QCI for Web sessions, downlink

VoIP If one is defined as QCI for VoIP, Retainability for VoIP

1.6 Detailed Separation of Observability in LTE RAN with

— Compensating KPIs with subtracting legacy PM counters of a KPI formula

37/1553-LZA 701 6014/1 Uen H | 2021-08-25 7

For list of filters supported by flexible counters, see Flexible Counters.

8 37/1553-LZA 701 6014/1 Uen H | 2021-08-25