Performance Analysis of Radio Resource Management

Strategies in UTRAN W-CDMA

J. Prez-Romero, O. Sallent, R. Agust, F. Casadevall
Universitat Politcnica de Catalunya
C/ Jordi Girona, 1 - 08034 Barcelona - Spain
E-mail : jorperez, sallent, ramon, ferranc @ tsc.upc.es
Abstract: This paper focuses on the performance evaluation of Radio Resource Management (RRM) for WCDMA. To this end, the framework to define RRM strategies is presented taking into account 3GPP
specifications. The simulation model to characterise system performance in such an environment is also provided
and finally some results are analysed for a selection of the presented strategies.
Topics: RRM, QoS provisioning, Admission Control, Congestion Control, Packet Scheduling

I.-INTRODUCTION
The key feature of third generation mobile systems will be the ability to deliver wideband and high bit rate
multimedia services alongside the traditional radio services such as voice, messaging and slow rate data. In
that context, UMTS (Universal Mobile Telecommunication System) will provide wideband mobile
multimedia services for the future mass market. The broad range of services expected to be supported
through these 3G networks can be divided into four Quality of Service (QoS) classes: conversational (e.g.
voice), streaming (e.g. video), interactive (e.g. www browsing) and background (e.g. e-mail). However, the
provision of such mobile multimedia services under QoS guarantees will not be possible without a proper
utilization of the air interface resources by means of Radio Resource Management (RRM) strategies that
ensure the target QoS, the planned coverage area and offer a high system capacity [1][2]. Such strategies
should deal with the peculiarities of the radio access technology, that in the UTRA FDD (UMTS Terrestrial
Radio Access Frequency Division Duplex) mode of UMTS is based on W-CDMA (Wideband Code
Division Multiple Access) [1]. One of the peculiarities of this access scheme is that it lacks from a constant
value for the maximum available capacity, since it is tightly coupled to the amount of interference in the air
interface. Therefore, RRM functions become crucial to manage this interference depending on the provided
services. Moreover, RRM strategies are not subject of standardisation, so that they can be a differentiation
issue among manufacturers and operators. Additionally, RRM functions can be implemented in many
different ways, this having an impact on the overall system efficiency and on the operator infrastructure cost,
so that definitively RRM strategies will play an important role in a mature UMTS scenario.
According to the 3GPP (3rd Generation Partnership Project) specifications, the radio interface of the
UTRAN (UMTS Terrestrial Radio Access Network) is layered into three protocol layers: the Physical Layer
(L1), the Data link Layer (L2) and the Network Layer (L3). Additionally, the layer 2 is split into two sublayers, the Radio Link Control (RLC) and the Medium Access Control (MAC). On the other hand, the RLC
and layer 3 protocols are partitioned in two planes, namely the User plane and the Control plane. In the
Control plane, Layer 3 is partitioned into sublayers where only the lowest sublayer, denoted as Radio
Resource Control (RRC), terminates in the UTRAN. Connections between RRC and MAC as well as RRC
and L1 provide local inter-layer control services and allow the RRC to control the configuration of the lower
layers. In the MAC layer, logical channels are mapped to transport channels. A transport channel defines the
way how traffic from logical channels is processed and sent to the physical layer. The smallest entity of
traffic that can be transmitted through a transport channel is a Transport Block (TB). Once in a certain period
of time, called Transmission Time Interval (TTI), a given number of TB will be delivered to the physical
layer in order to introduce some coding characteristics, interleaving and rate matching to the radio frame.
The set of specific attributes are referred as the Transport Format (TF) of the considered transport channel.
Note that the different number of TB transmitted in a TTI indicates that different instantaneous bit rates (i.e.
different Spreading Factors SF) are associated to different TF. Since the User Equipment (UE) may have
more than one transport channel simultaneously, whenever a combination of services is provided, the
Transport Format Combination (TFC) refers to the selected combination of TF. The network assigns a list of
allowed TFC to be used by the UE in what is referred as Transport Format Combination Set (TFCS) or
Transport Format Set (TFS) if only one transport channel is considered. Note also that this set defines the
maximum bit rate that can be used by the UE. Whenever a service is demanded by a certain UE, a Radio
Access Bearer (RAB) should be allocated to it. The RAB defines the way how transmissions in the radio
interface should be carried out in terms of type of transport channel, TTI, Transport Block size, possible TF
(i.e. TFS), MAC and RLC headers, as well as all the physical layer aspects such as channel coding,
interleaving, puncturing or slot formats.

71/1

Taking into account the constraints imposed by this radio interface architecture, the RRM functions are
responsible of taking decisions regarding the setting of the parameters such as TF, TFS, etc. as well as other
such as power level, code sequences, etc. RRM functions need to be consistent for both uplink and downlink,
although the different nature of these links introduces some differences in the followed approach. In
particular, RRM functions include:
o Admission control: It decides the admission or rejection of requests for setup and reconfiguration of radio

bearers. The admission control procedure should take into account the impact of handover users and
should be executed taking into account both uplink and downlink constraints.
o Congestion control: It faces situations in which the QoS guarantees are at risk due to the evolution of

system dynamics (mobility aspects, increase in interference, traffic variability, etc.).

o Short term RRM mechanisms: They are devoted to decide the suitable radio transmission parameters for

each connection in a reduced time scale and in a very dynamic way. Within these mechanisms the
following functions can be included:
- MAC algorithms: They are executed on a to decide the instantaneous Transport Format (or equivalently
instantaneous bit rate) to be applied in each TTI for a given RAB.
- Packet scheduling: It is responsible for scheduling non real time transmissions over shared channels. In
UTRA FDD this functionality manages the occupation over the DSCH (Downlink Shared CHannel)
- Power control: The purpose of this strategy is to optimise the mobile transmitted power (uplink) and the
base station transmitted power (downlink). To this end, power control is executed in two steps [1]:
- Inner loop power control: It is responsible of adjusting, on a fast time basis (i.e. each UTRA FDD
10 ms frame is subdivided into 15 slots each corresponding to a power control period), the
transmitted power in order to reach the receiver with the required Eb/No target .
- Outer loop power control: It is responsible of selecting a suitable Eb/No target depending on the
BLER (BLock Error Rate) or BER (Bit Error Rate) requirement. It operates on a slower time basis
than the inner loop power control, and adapts power control to changing environments.
o Code management: It is devoted to manage the downlink OVSF (Orthogonal Variable Spreading Factor)

code tree used to allocate physical channel orthogonality among different users [1].
o Handover control: The purpose of this strategy is to optimise the cell or set of cells (i.e. the Active Set

[1]) to which the mobile is connected.

The resulting decisions taken by RRM functions are executed by means of the radio bearer control
procedures, which define the signaling messages to be exchanged between the network and the UE.
Specifically, these messages are Radio Bearer Set-up, Physical Channel Reconfiguration and Transport
Channel Reconfiguration [1]. Figure 1 summarizes the framework described above for the RRM definition.
RRM OBJECTIVES:
1.
2.
3.
4.
5.

Maximize spectral efficiency

Prevent network congestion
Satisfy required QoS
Keep low signalling
Ensure planned coverage

RADIO BEARER CONTROL PROCEDURES:

1. Radio Bearer Set-up
2. Physical Channel Reconfiguration
3. Transport Channel Reconfiguration
Decisions taken by RRM functions are executed
through Radio Bearer Control Procedures

These objectives are achieved

through managing the following parameters
RRM FUNCTIONS:
Admission Control
Congestion Control
Short term RRM

MANAGED PARAMETERS:
1.
2.
3.
4.
5.

TFCS
TFC
Power
OVSF code
Active Set

These parameters are set by the

following RRM functions

MAC algorithms
Packet scheduling
Power control

Code Management
Handover control

Figure 1. Framework for RRM definition

The complete characterisation of the achieved system performance is a complex problem with many mutual
effects that depend on the specific solutions and parameters that have been considered for each of the RRM
strategies. To this end, the evaluation of the whole system is usually split in two simulation tools: link and
system level simulator. This paper mainly deals with the later, so that a multiuser, multicell and multiservice
scenario is considered. In turns, the physical layer performance is modelled from the outputs of the link level
simulator. The paper is organised as follows: section II presents the simulator that has been developed and
section III provides some sample results dealing with admission control for both uplink and downlink,
congestion control and packet scheduling over DSCH channel. Finally, section IV summarises conclusions.

71/2

II.- SIMULATION MODEL

The performance of the RRM algorithms has been evaluated through the use of a system level simulator that
has been developed by means of the OPNET simulation platform. It allows the support of a wide range of
RABs, traffic models as well as deployment scenarios. The simulator considers a representative service for
each of the four defined traffic classes, according to specific RABs selected from TS 34.108 [1]. The
mobility model and propagation models are defined in [3]. The physical layer characterisation is obtained
through a link level simulator [4] that feeds the system level simulator with the transport block BLER
statistics for each average (Eb/No). This characterisation includes a detailed simulation of all the processes
involved at the physical layer, such as channel estimation, antenna diversity, rate 1/3 turbo coding as well as
the 1500 Hz closed loop power control. Similarly, these link level results are also used to execute the outer
loop power control in the system level simulator (i.e. to compute required Eb/No given a BLER requirement).
From a functional point of view, the procedures to be considered in the system level simulator are reflected
in Figure 2(a). As an initial procedure, the network deployment module will be responsible for providing the
simulator user with a mechanism to introduce the position of the different base stations and mobile stations
as well as the other parameters characterizing the scenario to be evaluated. On the other hand, the RRM
module is the core of the simulator responsible for executing the analyzed RRM strategies. The RRM
module will act depending on the behavior of the mobile terminals in terms of traffic generation and
mobility. Regarding the mobility issues, the simulator will contain modules to implement the trajectories of
the terminals, to calculate the path loss to the base stations in the scenario and to decide the base stations in
the active set depending on the handover algorithms. Similarly, traffic generation models will be simulated
for each user depending on its corresponding service and the generated packets will be kept in buffers
waiting for transmission. The RRM module will decide when and how the packets are transmitted through
the radio interface. The power control mechanism will be responsible for determining the transmitted power
of each transmission to reach a certain Eb/No target. Depending on this power and the position of the
terminals the resulting Eb/No is evaluated for each transmission. Finally, the interaction with the off-line link
level simulator results will decide the successful and erroneous transmissions. The buffers will be updated
accordingly depending on the result for each transmission and on the availability of retransmissions.

TRAJECTORIES

PATH LOSS
CALCULATIONS
BS-BS
BS-MS
MS-MS

NETWORK
DEPLOYMENT

TRAFFIC

HANDOVER
ALGORITHM

BUFFERING

Admission Control
Congestion Control
Short term RRM
Code Management

POWER
CONTROL

RRM

STATISTICS FROM
LINK LEVEL
SIMULATOR

SIGNAL AND INTERFERENCE

EVALUATION
(Eb/No Calculations)

(a)
(b)
Figure 2. Functional Simulator Architecture (a) and Network Model (b)

Figure 2(b) shows a network model representing a possible scenario under test for the system level simulator.
Particularly, the following nodes have been developed to implement the previously described functions: the
RNC node, which simulates the behavior of the Radio Network Controller and particularly deals with the
simulator control and with the RRM strategies, the BS node, the UE node, which simulates the behavior of a
User Equipment in terms of mobility, traffic generation and radio transmission functions, and the Fixed
Network node, which acts as the generation source for downlink traffic. Furthermore, the Rest of Users
node simulates a high number of UEs without having to locate all of them in the network model. The only
functional difference between a UE node and a user in the Rest of Users node is that the first follows a userdefined trajectory while the second makes use of a mobility model.
III.- PERFORMANCE RESULTS
This section provides an overview of some of the results that have been obtained with the previously
described simulation tool for the different RRM strategies. Since RRM for W-CDMA is a complex problem

71/3

with many mixing effects, this section is organised as a collection of different case studies each one focusing
on a given RRM function either in uplink or in downlink. This will allow to devise the main parameters and
factors influencing on the achieved performance as well as the key elements of the different RRM strategies.
The results presented here consider an UL/DL 64 kb/s CBR videophone service (representative of the
conversational service class) and a WWW service (representative of the interactive class) with a maximum
bit rate of 64 kb/s in the UL and 256 kb/s in the DL. Simulation parameters are summarised in Table 1.
Table 1. Simulation parameters
2.25 km x 2.25 km
10 ms
Frame duration
BS parameters
UE parameters
Cell radius
500 m
Maximum transmitted power
21 dBm
Cell type
Omnidirectional
Minimum transmitted power
-44 dBm
Maximum transmitted power
43 dBm
Thermal noise
-100 dBm
Thermal noise
-106 dBm
Mobile speed
50 km/h
Common Control Channels Power
30 dBm
DL Orthogonality factor
0.4
Propagation model (macrocellular)
Shadowing deviation
10 dB
Shadowing decorrelation length
20 m
Handover parameters (conversational)
Handover parameters (interactive)
Active Set (AS) maximum size
2
Active Set maximum size
1
Threshold to enter AS
3 dB
Replacement hysteresis
1 dB
Hysteresis to enter AS
1 dB
Time to Trigger Handover
0.5s
Replacement hysteresis
1 dB
Time to Trigger Handover
0.5s
Traffic model (interactive) [3]
Traffic model (conversational)
Call duration
120s (exponential)
Number of pages per session
5 (geometrical)
Offered bit rate
64 kb/s (CBR)
Reading time between pages
30 s (exponential)
Activity factor
1
Number of packets per page
25 (geometrical)
Call rate
29 calls/h/user
Time between packets
0.0625 s (DL)
(Poisson arrivals)
(exponential)
0.125 s (UL)
Packet length
Pareto distributed
(=1.1, k=81.5 bytes,
m=6000 bytes)
Session rate
25 sessions/h/user
QoS parameters (conversational)
QoS parameters (interactive)
BLER target
1%
Average page delay
<4s
Dropping probability
<1%
Dropping probability
<1%
Dropping condition
1s below Eb/No target
Dropping condition
Lack of codes in the
or lack of codes in the
new cell during HO
new cell during HO
Scenario size

3.84 Mcps

Chip Rate W

III.1.- Uplink admission control

In W-CDMA there is a close relationship between coverage and capacity, since in such an interference
limited scenario, the power required by a given connection is directly related to the number of users that are
simultaneously transmitting. Specifically, the power to be transmitted by each UE of a given cell depends on
the cell uplink load factor UL and the path loss Lp(di) (including shadowing) at distance d i as [5]:
1
1 UL
= L p (d i )
W
+1
Eb

Rb ,i
N o i
PN

PT ,i

i=1..n

(1)

where n is the number of users transmitting simultaneously at a given moment, Rb,i is the i-th user
instantaneous bit rate, W is the total bandwidth after spreading and PN is the thermal noise power. The uplink
load factor provides the theoretical spectral efficiency of a W-CDMA cell and is given by [7]:
(2)
n

UL = 1 +
P
R i =1

W
Eb

No

Rb ,i
i

+1

PR +
PR + + PN

71/4