INTL JOURNAL OF ELECTRONICS AND TELECOMMUNICATIONS, 2020, VOL. 66, NO. 1, PP.

211-216
Manuscript received November 15, 2019; revised January, 2020. DOI: 10.24425/ijet.2020.131865

Performance Analysis of Dynamic PUCCH

Allocation Algorithm in LTE Network
Maciej Czerniecki, Jacek Wszołek, Marek Sikora and Wiesław Ludwin

Abstract—The aim of the presented paper was to verify the The channel bandwidth defines the number of available RBs.
impact of Dynamic PUCCH Resource Allocation Algorithm of the When the LTE system serves 20 MHz channel, it has 100 RBs
LTE cellular system on the maximum uplink cell throughput and available for schedule. In order to maximize the number of
call setup success rate - CSSR. Paper includes the laboratory contiguous PRBs that can be allocated for the PUSCH
testbed description and presents the results of an experiment conveying user traffic, PUCCH resources are located on the
confirming the improvement of both key performance indicators extreme ends of the bandwidth (Fig. 1). (Note that up to Release
KPIs. Apart from the presentation of the Dynamic PUCCH
Resource allocation algorithm, the paper also includes a
10 only continuous bandwidth could be allocated to terminals).
description of legacy LTE uplink (PUCCH and PUSCH) channels To provide frequency diversity gain for the PUCCH
dimensioning process thus filling the gap of such a tutorial in the transmissions UEs hop between the bandwidth edges (intra
available literature. subframe hopping) [5].

Keywords—LTE, optimization, planning, PUCCH

dimensioning, network planning, network optimization, Self-
Organizing Networks, SON, Dynamic PUCCH allocation, Static
PUCCH allocation

I. INTRODUCTION

P LANNING and optimization process of the LTE cellular

network [1], apart from many steps, requires also suitable
PUCCH (Physical Uplink Control CHannel) dimensioning [2],
[3]. PUCCH dimensioning process [2] relies on assigning the
appropriate number of radio resource blocks (RB) in the uplink
for PUCCH control channel and on the configuration of the
Fig. 1. PUCCH resource in time and frequency domain.
maximum frequency (periodicity) of SR (Scheduling Request)
messages and CSI (Channel State Information) reports [4].
Each UE in the RRC Connected mode (connected to the LTE
During PUCCH planning process operator can also assign
network and transmitting data) is obliged to send L1/L2 control
dedicated PUCCH RBs for special purposes like for sending
signalling on PUCCH periodically [6]. The configuration of a
HARQ (ACK/NACK) and SR only [4]. While the PUCCH given eNodeB (base station) determines the periodicity of
configuration is vendor dependent, the standardized sending both SR messages and CSI reports. The configuration
configuration results are broadcasted with the aid of System is most often the result of an optimization process [1], [2].
Information Block Type 2 (SIB2) [4]. Frequent and accurate reporting of CSI and the ability to send
The amount of radio resources in the LTE uplink depends on SR messages frequently requires a large amount of radio
the system bandwidth, which can vary from 1.4 up to 20 MHz resources on the PUCCH channel, which limits the maximal
and division duplex strategy, which can be either TDD (Time number of active users, but on the other hand, improves the
Division Duplex) or FDD (Frequency Division Duplex) [5]. All quality of services offered by the LTE network [1], [2].
LTEs uplink radio resources can be allocated either for PUCCH One of the most important parts of the PUCCH L1/L2
[3] or for PUSCH (Physical Uplink Shared Channel) [3]. signalling is transmission of a Scheduling Request SR. The RRC
PUCCH is a control channel dedicated to carry UE specific connected UE, which is willing to transmit data, sends
L1/L2 signalling including scheduling request, hybrid ARQ scheduling request indicator (SRI) to the eNodeB [4]. If the base
messages and Channel State Information (CSI) which can be station has enough radio resources, it replies to a scheduling
composed of Channel Quality Indicators (CQI), Rank Indicator request by giving the mobile uplink grant. Terminals are also
(RI) and Precoding Matrices Indicators (PMI) [4]. On the other having other possibilities to get uplink resources like sending
hand PUSCH is a channel which carries uplink user data. BSR (Buffer Status Report) MAC control signalling [6] or they
can just receive some proactive uplink grants [7], [8].

The work of Jacek Wszołek, Marek Sikora and Wiesław Ludwin was Jacek Wszołek, Marek Sikora and Wiesław Ludwin are with AGH
supported by the Polish Ministry of Science and Higher Education with the University of Science and Technology (e-mail: [email protected],
subvention funds of the Faculty of Computer Science, Electronics and [email protected], [email protected]).
Telecommunications of AGH University Maciej Czerniecki is with Nokia Solutions and Networks, Kraków, Poland
(e-mail: [email protected])

© The Author(s). This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0,
https://creativecommons.org/licenses/by/4.0/), which permits use, distribution, and reproduction in any medium, provided that the Article is properly cited.
212 M. CZERNIECKI, J. WSZOŁEK, M. SIKORA, W. LUDWIN

Nevertheless during PUCCH dimensioning process SR service quality level [1], [2]. As a result of the PUCCH planning
resources must be reserved for each connected user [1], [2], [4]. and optimization process following parameters are determined:
Each UE, which is in the RRC Connected state, must also have
reserved resources on the PUCCH channel for sending Hybrid • The number of RB assigned to PUCCH (PUCCH
ARQ (HARQ). It is absolutely necessary for positive or area)
negative acknowledgement of downlink traffic. Therefore, the • SR Periodicities
maximal number of connected UEs in the cell depends on the
amount of resources dedicated to PUCCH. • CSI Periodicities
High SR periodicity increases the likelihood of more frequent The process of dimensioning the PUCCH channel is further
allocation of radio resources to a selected user, and thus reduces complicated by periodic fluctuations in LTE network traffic
the delay. In turn, more accurate reporting of the radio channel load. For example, during busy hours (BH) it is important to
state increases the efficiency of the AMC (Adaptive Modulation have a possibility to serve as many users as possible and
and Coding) [9], [10], [11] and CAPS (Channel Aware Packet decrease the rejection probability. Otherwise Call Setup Success
Scheduling) algorithms [12], [13], which translates into reduced Rate CSSR KPI can become significantly degraded. Therefore,
packet loss and increased spectral efficiency [1]. from the maximal number of connected UEs point of view, it is
As a rule of thumb, L1/L2 control signalling messages are sent worth allocating more resources for the PUCCH channel.
via the PUCCH channel. Despite the fact that, clustered SC- On the other hand, after busy hours, when the traffic load is
FDMA technique for UL transmission, which allows non- small, a large amount of resources allocated for the PUCCH
contiguous groups (clusters) [3] of subcarriers to be allocated to channel translates into limiting the maximum uplink bandwidth.
a single terminal in one TTI (Transmission Time Interval), were In this situation, a better solution is to limit the amount of
introduced in Release 10, even today most terminals do not resources allocated to the PUCCH channel.
support that feature. Therefore most UEs are not capable of Until now, the LTE network operator, could periodically
sending both PUCCH and PUSCH simultaneously. When the assign the number of RBs for the PUCCH channel and
UE is scheduled an uplink grant and in a meantime is also determine the periodicity of SR and CSI messages. There is a
obligated to transmit L1/L2 control signalling it can send them possibility to repeat this process very often, although in practice,
both using the PUSCH channel. However, from this paper’s updates in the PUCCH configuration were performed relatively
point of view, it does not matter much, because regardless of rarely, almost only when selected KPIs (e.g. Call Setup Success
whether these messages are sent via PUCCH or PUSCH, they Rate - CSSR) exceeded the quality threshold. Static PUCCH
occupy radio resources and increase a signalling overhead. dimensioning is a complex process. Moreover, it is not possible
When optimizing selected cell for the maximum uplink to find optimal PUCCH area because it depends on number of
throughput [1], it is best to allocate the maximum number of UEs in RRC connected state which fluctuates in time.
RBs to the PUSCH channel, thus limiting the signalling
overhead (PUCCH) to the necessary minimum. On the other II. STATIC PUCCH PLANNING AND DIMENSIONING
hand, limiting the number of radio resources allocated for the
PUCCH channel translates into a reduction in the maximal As described above, dimensioning of the PUCCH area is very
number of users in RRC Connected mode. important process. If too small amount of resources is dedicated
Both the maximum number of users in the RRC Connected to PUCCH a poor call setup success rate is usually observed. On
mode and the maximum uplink throughput are Key Performance the other hand, overestimation of PUCCH implicate higher
Indicators (KPI) of the LTE network. signalling overhead which causes lower user data throughput.
The paper is organized as follows. A short introduction to Therefore in order to achieve appropriate balance between
PUCCH Dimensioning is presented in Section I. Section II control and data traffic resources the PUCCH size needs to be
contains a static PUCCH planning and optimization basics. In planned.
Section III, a dynamic PUCCH allocation algorithm is TABLE I
PUCCH FORMATS
described. Section IV contains description of test environment
while the analysis of the obtained results are shown in Section PUCCH
Usage Bits Modulation
V. The paper is finalized with a Summary and Conclusions. Format
1 Scheduling Request N/A ON/OFF
I. PUCCH DIMENSIONING Keying
1a 1-bit HARQ-ACK & SR 1 BPSK
The process of dimensioning and optimization of the PUCCH 1b 2-bit HARQ-ACK & SR 2 QPSK
channel is a trade-off among the maximum uplink throughput, 2 CQI/PMI or RI report 20 QPSK
the maximal number of RRC Connected UEs and the quality of 2a CQI/PMI or RI report & 1- 21 QPSK+BPSK
service measured by QoS indicators, such as latency or bit HARQ-ACK
2b CQI/PMI or RI report & 2- 22 QPSK+BPSK
throughput. Therefore, from the network optimization point of bit HARQ-ACK
view, the PUCCH region needs to be planned in order to get an
appropriate balance between amount of radio resources
dedicated to signalling and user data. Optimal PUCCH area The PUCCH is located at the lower and upper band edges of
configuration should reduce probability of UE rejection caused the UL system bandwidth (Fig. 1). RB carrying CQIs (PUCCH
by the lack of uplink L1/L2 signalling resources, provide format 2/2a/2b - Table I) are located at the outermost part of the
maximum possible user data uplink throughput and assure high uplink bandwidth, while the SR and the HARQ ACK/NACKs
PERFORMANCE ANALYSIS OF DYNAMIC PUCCH ALLOCATION ALGORITHM IN LTE NETWORK 213

PUCCH RBs (format 1/1a/1b - Table I) are located at the RBs should be assigned for format 2/2a/2b. The PUCCH format
innermost PUCCH resource blocks [3]. 2.x capacity is calculated by:
A. PUCCH Format 1/1a/1b
1[RB]*6[CyclicShifts]*[CQIPeriodicity]*nCQIRB (2)
When using PUCCH formats 1, 1a and 1b (Table I), the
mobile modulates the bits onto one symbol, using BPSK for a
one-bit acknowledgement (PUCCH format 1a) and QPSK for a For example, to serve 840 UEs in a cell with
two-bit acknowledgement (PUCCH format 1b). It then spreads CQIperiodicity = 20ms, the following number of RB for format
the information in the time domain using the orthogonal 2.x is needed:
sequence, usually across 4 symbols, (3 symbols in slots which 840/(6 UEs per RB * 20 ms) = 7 RB (3)
support an SRS - Sounding Reference Signal). The mobile then
spreads the information across 12 sub-carriers in the frequency III. DYNAMIC PUCCH ALLOCATION ALGORITHM
domain using the cyclic shift. Having 3 orthogonal sequences
In real network, applying frequent updates in the PUCCH
and 12 cyclic shifts it is theoretically possible to schedule
12x3 = 36 UEs in a single RB. However it practice it is not region manually as well as introducing changes in periodicities
recommended to schedule more than 18 UEs in one PUCCH RB of SR and CSI in such a way is almost unfeasible. Hence
dynamic PUCCH algorithm was developed, that enables the
[5].
modification of the number of physical resource blocks (PRBs)
TABLE II allocated to the physical uplink control channel (PUCCH).
MAX NUMBER OF UES SCHEDULED IN A SINGLE PUCCH RB
Adaptation is made proportionally to the number of UEs in
PUCCH Multiplexing RRC connected state in a cell (cell load). Using appropriate
Purpose Bits parameters operator can define both, PUCCH and PUSCH
Format capacity UE/RB
1 Scheduling Request N/A 36, 181, 12 channels adaptations areas (Fig. 2) and have control on the range
1a 1-bit HARQ-ACK & SR 1 36, 181, 12 of dynamic PUCCH allocation algorithm.
1b 2-bit HARQ-ACK & SR 2 36, 181, 13
2 CQI/PMI or RI report 20 36, 181, 13
2a CQI/PMI or RI report & 1- 21 12, 61, 4
PUCCH AREA

bit HARQ-ACK

MAXIMUM POSSIBLE PUCCH CONFIGURATION

2b CQI/PMI or RI report & 2- 22 12, 61, 4
bit HARQ-ACK
1
recommended maximal value.

MINIMAL PUCCH CONFIGURATION

FIRST PUCCH CONFIGURATION
The PUCCH format 1/1a/1b capacity is calculated as follows:
PUSCH AREA

MultiplexingCapacity(18)*SRPeriodicity (1)

Where value of MultiplexingCapacity is taken from Tab. 2 PUSCH

and the SRPeriodicity defines the admitted Scheduling Request
PRACH

periodicity in the cell. For example, with SRPeriodicity=20ms

scheduling capacity is:
PUCCH AREA

18 [users per RB] * 20 [ms] = 360 Users.

B. PUCCH Format 2/2a/2b

Format 2/2a/2b is capable of conveying 5 symbols per slot

which gives 20 coded bits plus ACK/NACK per subframe. The
PUCCH resource blocks for PUCCH format 1/1a/1b and format Fig. 2. PUCCH adaptation areas.
2/2a/2b can be allocated either separately or mixed according to
LTE specification [3]. Moreover, algorithm can adjust scheduling request (SR) and
The PUCCH format 2/2a/2b carries CQI/ PMI and HARQ channel state information (CSI) periodicities dynamically.
ACK/NACKs. Every UE with established SRB1 (Signalling When using dynamic PUCCH resource allocation algorithm,
Radio Bearer 1) is obligated to send CQI reports periodically adaptation of PRBs assigned to PUCCH is more efficient and
[5], which means that every UE requires format 2/2a/2b swift because PUCCH size is adequate to the number of UEs
resource assignment. Table II presents PUCCH formats 2.x, connected to the cell.
multiplexing capacity. It is possible to allocate 4, 6 or 12 UE per When governed by the dynamic allocation algorithm, the
single RB, however it is not recommended to exceed 6 UE/RB. PUCCH size can be expanded or compressed. In the dynamic
In real life network total capacity for format 2.x in a cell is algorithm, two procedures have been implemented. First
defined by the number of dedicated PRBs, cyclic shifts (6) and procedure defined as PEP (PUCCH Expansion Procedure) is
CQI periodicity. In the opposite of format 1.x for format 2 responsible for expanding PUCCH region. Oppositely to PEP,
nCQIRB parameter (Number of Physical Resource Blocks that second procedure named PCP (PUCCH Compression
are reserved for PUCCH formats 2/2a/2b) defines how many Procedure) reduces PUCCH region.
214 M. CZERNIECKI, J. WSZOŁEK, M. SIKORA, W. LUDWIN

During PEP procedure, new RBs are assigned to PUCCH In some specific cases, after PUCCH reconfiguration, the
channel, with the same periodicities of SR and CSI. It allows to number of UEs in RRC Connected state can change
increase cell capacity significantly, which is measured by immediately. Such situation requires dynamic PUCCH
maximum UEs in RRC Connected state per cell. It is possible to algorithm response. If PUCCH expansion is triggered and the
calculate analytically the PUCCH capacity with specified number of SIB modification per interval is higher than SIB
number of assigned RBs and periodicities of SR and CQI. update rate threshold, PUCCH size is reconfigured to the
The PEP procedure can be controlled by two parameters: maximum allowable value. On the other hand, if compression
procedure is triggered and the number of SIB modification per
• Upper threshold UEs in RRC connected, interval is higher than SIB update rate threshold, PUCCH
• Countdown timer SIB modification by PUCCH compression procedure is aborted and put off to the next
expansion. countdown expiration.
The first parameter is defined as a percentage of the number IV. THE TESTBED ENVIRONMENT
of UEs in RRC connected state. It defines threshold which
triggers expansion, so adding next RBs to the PUCCH channel. Experiments were carried out in a laboratory environment
By default, it is set to 80% of maximum number of UEs which which is depicted in Fig. 3. The laboratory environment
can be served (PUCCH capacity) for a given PUCCH consisted of the ET emulator, a hardware emulating multiple
configuration. UEs behaviour, the real eNodeB hardware of the LTE system,
To avoid ping-pong effect between PUCCH configurations the EPC (Evolved Packet Core) emulator and the TGRs Traffic
which means contiguous reconfiguration of the number of RBs Generators. The Multi-UE-Emulator was attached to both the
assigned to PUCCH region, the second parameter was uplink traffic generator and the eNB. The eNodeB under the test
introduced. It is defined as a counter which defines how long it was configured as a three sectored cell serving a bandwidth of
takes to make next PEP procedure. Default value of the 20 MHz per each sector. Each sector was configured for
countdown timer for changing the number of RB allocated for dynamic open loop 2x2 MIMO without carrier aggregation and
PUCCH is 10 minutes. enhanced intercell interference cancelation (eICIC) features.
The base station informs UEs about PUCCH configuration The EPC part including MME (Mobility Management Entity),
via SIB2 (System Information Block Type 2) [4], which contains PGW (Packet Data Network Gateway), SGW (Serving
radio resource configuration information that is common for all Gateway) was emulated using Nokia proprietary software EPC
UEs, and is periodically broadcasted by base station. Every UE emulator configured to support large number of UEs (2520
which is in RRC connected or idle state must receive and process UEs). In all experiments performed during the test campaign,
all SIBs information [4], [5]. the constantly full-buffer model of UDP data stream was used.
The PCP procedure compress radio resources assigned to The input traffic stream for the whole group of UEs was
PUCCH. The mechanism is similar to PEP procedure in terms generated by TGRs (Traffic Generators), serving the traffic to
of corresponding steering parameters. In the opposite to PEP the the ET emulator for the uplink direction. The role of the ET
PUCCH compression procedure needs much more attention. emulator was to evenly distribute traffic streams among N
However, it needs to be pointed that there are some potential independent radio transmissions. As a results of the distribution
complications during PUCCH region compression like: N different virtual EUs were created each transmitting a full-
buffer UDP traffic on the default radio bearer (best effort
• Connection rejection caused by lack of radio traffic). The ET emulator was connected with the eNodeB via
resources for signalling for some already connected waveguide with a 40 dB attenuator, which has to reduce signal
UEs, power to a correct level. In order to make the results free from
the influence of the radio propagation environment no channel
• Risk of unsuccessful relocation during PUCCH
emulation were used in the test setup.
resource reconfiguration event
Therefore, in the PCP procedure steering parameters has Traffic Traffic
much higher values than in the PEP procedure. The Countdown Generator UL Generator DL
timer SIB modification by PUCCH compression is set to 60
minutes by default and the number of UEs requiring
reconfiguration must be limited to avoid unnecessary
repositioning. Similarly the Lower threshold UEs in RRC
connected by default is set to 50%.
The algorithm reacts when number of UE change, every 40dB ATT EPC
PUCCH configuration can support certain number of UEs in
RRC connected state which is calculated by the eNodeB. During eNodeB
expansion when Upper threshold UEs in RRC connected Multi-UE Emulator
exceed, the eNodeB decides to perform reconfiguration and add
next RBs for the PUCCH channel. The number of RBs which Fig. 3. Simplified testbed environment schematic
are added to expand PUCCH region is not constant and depends
on the eNodeB configuration (MIMO, Carrier Aggregation, V. THE EXPERIMENT
eICIC (enhanced Inter-Cell Interference Coordination), etc.) The experiment was conducted with an eNodeB configured
[14]. for 20 MHz bandwidth. The aim of the experiment was to
PERFORMANCE ANALYSIS OF DYNAMIC PUCCH ALLOCATION ALGORITHM IN LTE NETWORK 215

evaluate the influence of dynamic PUCCH allocation algorithm The maximum cell uplink throughput achieved during
on the basic cell's uplink performance indicators including experiments is presented in the Fig 4. After quick analysis of the
maximum uplink throughput (MaxCellULThroughput) and call Fig. 4 it can concluded that the maximum achievable uplink
setup success rate CSSR. throughput decreases with the growing number of UEs attached
Four different base station configurations were used during to the eNB. This phenomenon is a clear consequence of
the experiment. In the first configuration, the allocation of the increasing amount of resources devoted to serving signalling
PUCCH cell resources was controlled by the algorithm which needs of active UEs.
dynamically adjusted the periodicity of the SR and CQI. The The variability of maximum achievable uplink throughput is
rest three static configurations are treated as a baseline. Detailed most significant for the first scenario where the number of
information concerning baseline configurations are collected in resource blocks designated to PUCCH channel was governed by
the adaptive algorithm. In case of the three baseline scenarios
Table III.
the decrease in maximum uplink throughput was significantly
TABLE III smaller due to the constant number of resource blocks
BASELINE CONFIGURATIONS WITH STATIC PUCCH ALLOCATION designated to form a PUCCH channel. The number of resource
blocks devoted for PUCCH channel in baseline scenarios
PUCCH allocation Number of PRB allocated remained constant despite of variability in a traffic load. As a
Number
algorithm to PUCCH
consequence for the low values of the traffic offered by UEs,
1 Dynamic PUCCH 4-8 most of the signalling channel resources were wasted. On the
2 SR: 20 ms, CQI: 80 ms 8
3 SR: 40 ms, CQI: 40 ms 10 other hand when the traffic offered by UEs approaches a cell
4 SR: 20 ms, CQI: 40 ms 10 capacity, not only the whole number of resource blocks devoted
to the PUCCH channel are utilised, but additionally some of the
PUSCH resource blocks are used to convey the signalling
During all experiments UEs were configured to offer traffic information.
exceeding the total theoretical capacity of the eNodeB Closer look at the Fig. 4 reveals that highest increase of the
regardless of number of the UEs (full-buffer traffic model). The uplink throughput can be observed for the case of dynamic
experiment was divided into six stages for each of which, in all PUCCH resource allocation. The lower the number of the UEs
four configurations, presented in Table III, the number of the the higher was the gain in maximum uplink throughput. For the
UEs attached do the eNodeB was increased. The number of the 50 terminals case, depending on configuration the gain varied
attached terminals was increased from 50 through 100, 300, between 5% and 8% which was caused by reallocation of the
500, 700 up to 840, where 840 was a maximum number of unutilized resource blocks designated formerly to PUCCH to
terminals supported by the eNodeB. the traffic channel PUSCH.
During the experiment the value of the maximum uplink cell The number of the resource blocks allocated for PUCCH by
throughput averaged over a 10 seconds intervals and the number the dynamic PUCCH allocation algorithm were collected in
of resource block reserved for the PUCCH by the dynamic Table IV. The case of the 50 terminals is an interesting example
resource allocation algorithm were collected. Additionally of the PUCCH resource allocation. For the case of the static
during the last stage with 840 UEs attached the CSSR was also PUCCH resource allocation the number of 4 RBs allocated for
monitored. the uplink signalling would be connected with strong risk of the
drastic reduction of the cell capacity during the traffic busy
hours. Such a conservative allocation of the PUCCH resources
Dynamic PUCCH would in the static allocation case limit the maximum number
50 of active UEs in a cell to approximately 100. With the dynamic
SR: 20ms, CQI: 80ms
PUCCH resource allocation the risk of reduction of the cell
SR: 40ms, CQI: 40ms capacity can be avoided since with the increase in number of the
Max Cell UL Throughput [Mbit/s]

49
active UEs the dynamic PUCCH allocation algorithm will
SR: 20ms, CQI: 40ms instantly increase the number of resource blocks designated to
48 PUCCH thus avoiding a risk of the cell blockage.

47 TABLE IV
NUMBER OF RESOURCE BLOCKS DESIGNATED TO PUCCH CHANNEL

46
Number of RBs allocated for PUCCH
Number
45 of active
Static PUCCH
UEs in Dynamic
the cell PUCCH SR: 20ms, SR: 40ms, SR: 20ms,
44 CQI: 80ms
CQI: 40ms CQI: 40ms
50 4 10 10 8
43 100 6 10 10 8
50 100 300 500 600 840 300 6 10 10 8
500 6 10 10 8
Number of active UEs
700 8 10 10 8
840 8 10 10 8
Fig. 4. Maximum cell uplink throughput
216 M. CZERNIECKI, J. WSZOŁEK, M. SIKORA, W. LUDWIN

VI. THE SUMMARY RERENCES

The Table V summarizes the selected values of counters and [1] Ajay R. Mishra, Fundamentals of Network Planning and Optimization
2G/3G/4G: Evolution to 5G, John Wiley & Sons, 27 July 2018.
performance indicators collected as part of research experiments
[2] Xincheng Zhang, LTE Optimization Engineering Handbook, John Wiley
for the eNodeB LTE base station serving up to 840 active users. & Sons, 4 Jan 2018, ch. 4.3, pp 144-152
For comparative purposes, two scenarios were considered: a [3] Physical channels and modulation (Release 14), 3GPP TS 36.211
static and dynamic PUCCH configuration where the last one V14.9.0. 2018
[4] Evolved Universal Terrestrial Radio Access (E-UTRA); Radio Resource
was governed by the dynamic PUCCH allocation algorithm.
Control (RRC); Protocol specification (Release 14), 3GPP TS 36.331
TABLE V V14.12.0. 2019
SELECTED VALUES OF COUNTERS AND KPIS FOR THE ENODEB SERVING UP TO [5] Physical layer procedures (Release 14), 3GPP TS 36.213 V14.9.0 2018.
840 ACTIVE UES FOR STATIC AND DYNAMIC PUCCH RESOURCE ALLOCATION [6] Evolved Universal Terrestrial Radio Access (E-UTRA); Medium Access
Control (MAC) protocol specification (Release 14), 3GPP TS 36.321
KPI/ Dynamic Static V14.11.0. 2019
Counter name PUCCH PUCCH [7] J. Brown and J. Y. Khan, "Predictive resource allocation in the LTE uplink
for event based M2M applications," 2013 IEEE International Conference
RRC CSSR 100% 42.86%
on Communications Workshops (ICC), Budapest, 2013, pp. 95-100.
SIGN_CONN_ESTAB_FAIL_PUCCH 0 480
[8] J. Brown and J. Y. Khan, "A Predictive Resource Allocation Algorithm in
SIGN_CONN_ESTAB_COMP 840 360
the LTE Uplink for Event Based M2M Applications," in IEEE
Transactions on Mobile Computing, vol. 14, no. 12, pp. 2433-2446, 1 Dec.
2015
In case of the static PUCCH allocation the eNodeB under the [9] S. H. O. Salih and M. M. A. Suliman, "Implementation of adaptive
test was able to serve up to 360 active UEs which reflected the modulation and coding techniques using Matlab," Proceedings ELMAR-
value of the call success rate counter of 42,86% being totally 2011, Zadar, 2011, pp. 137-139.
unacceptable for the commercial network. [10] S. El Makhtari, M. Moussaoui and H. Samadi, "Performance evaluation
of AMC turbo coded OFDM for 3GPP Long Term Evolution downlink
Contrary to the static case the introduction of the dynamic system," 2012 International Conference on Multimedia Computing and
PUCCH resource allocation algorithm enabled the eNodeB to Systems, Tangier, 2012, pp. 490-494.
serve up to 840 active UEs. It's worth noting that such a large [11] S. Avallone, N. Pasquino, S. Zinno and D. Casillo, "Experimental
number of active UEs subtly decreased the uplink throughput characterization of lte adaptive modulation and coding scheme under
actual operating conditions," 2017 IEEE International Workshop on
comparing to the static allocation case. Measurement and Networking (M&N), Naples, 2017, pp. 1-6.
The decrease in maximum achievable uplink throughput was [12] R. Sandanalaksmi, K. Manivanan, S. Manikandan, R. Barathi and D.
caused by the increase of the number of RBs designated for the Devanathan, "Fair channel aware packet scheduling algorithm for fast UL
PUCCH signalling purposes, resulting from the increased HARQ in UTRAN LTE," 2009 International Conference on Control,
Automation, Communication and Energy Conservation, Perundurai,
number of active users. Tamilnadu, 2009, pp. 1-5.
[13] Y. Lin and G. Yue, "Channel-Adapted and Buffer-Aware Packet
VII. CONCLUSIONS Scheduling in LTE Wireless Communication System," 2008 4th
International Conference on Wireless Communications, Networking and
Evaluating dynamic PUCCH allocation algorithm from the Mobile Computing, Dalian, 2008, pp. 1-4.
mobile network operator perspective, it is an especially [14] Evolved Universal Terrestrial Radio Access (E-UTRA) and Evolved
beneficial technology for those operators who are reaching the Universal Terrestrial Radio Access Network (E-UTRAN); Overall
description; Stage 2, 3GPP TS 36.300 V14.11.0. 2019
uplink capacity limits in their networks. Introduction of [15] Srikant R., Ying Lei. 2013. “Communication Networks: An Optimization,
presented algorithm allows operators to increase the uplink Control, and Stochastic Networks Perspective”. Cambridge University
throughput by up to 8%, which is especially important now, Press.
when the demands for uplink are growing significantly. This is [16] Ericsson Mobility Report, November 2019
[17] F. B. Mismar, J. Choi and B. L. Evans, "A Framework for Automated
caused mostly by social networking which has a high uplink Cellular Network Tuning With Reinforcement Learning," in IEEE
demands as users actively create content [16]. Transactions on Communications, vol. 67, no. 10, pp. 7152-7167, Oct.
Moreover, dynamic PUCCH algorithm is a part of SON (Self- 2019. doi: 10.1109/TCOMM.2019.2926715
Organizing Networks) [17], [18], which allows operator to [18] E. Aqeeli, A. Moubayed and A. Shami, "Dynamic SON-Enabled Location
Management in LTE Networks," in IEEE Transactions on Mobile
optimize OPEX (Operating Expenses). It also simplifies the Computing, vol. 17, no. 7, pp. 1511-1523, 1 July 2018. doi:
network dimensioning process and therefore decrease TTM 10.1109/TMC.2017.2775221
(Time to Market) as well.