The main advantages of second-generation mobile systems (e.g. Global System for Mobile commu- nication (GSM)) over their analog predecessors are higher capacity and lower battery consumption. However, the second-generation has retained the circuit-switched legacy of analog network, which is designed to carry voice traffic [Oli99]. Moreover, data rates in GSM were limited (e.g. 14.4 Kbps for voice traffic). Presently, the capabilities of second-generation systems is extended by adding some multimedia capabilities, such as the support for high bit rates and the introduction of packet data/IP access [DGNS98]. These improvements have been introduced by new technologies such as HSCSD (High Speed Circuit Switched Data), GPRS (General Packet Radio Service) or EDGE (Enhanced Data Rates for GSM Evolution) and allowed the increase of data rates up to 384 Kbps [LGT99].
In order to create a unified telecommunications system with multimedia capabilities, the Inter- national Telecommunication Union (ITU) has defined a framework for third generation telecom- munications systems, called IMT-2000 (International Mobile Telecommunications-2000). Within the IMT-2000 framework, the Universal Mobile Telecommunications System (UMTS) has been developed and standardized by the Third Generation Partnership Project (3GPP).
1.1.1 Major Features of UMTS
One of UMTS objectives is to provide a global coverage for all types of users (figure1.1); therefore, RRM procedures and radio access techniques are developed to cover users in indoor, outdoor, urban and rural environments with mobility ranging form pedestrian up to vehicular with very high speed. Moreover, UMTS is intended to offer high quality multimedia services. These types of services require high data rates; therefore, the radio interface of UMTS is intended to allow up to 10.2 Mbps data rates [UMT04] in indoors and micro-cells. Furthermore, some multimedia services are not very sensitive to delay, and thus a packet-switched legacy can optimize the use of scarce radio resource units. This approach is considered by UMTS architecture (e.g. an all-IP architecture is proposed in Release 5 and 6 of 3GPP specifications). UMTS architecture allows also users to
1.1 Universal Mobile Telecommunications System 3
Table 1.1: UMTS service classes
Traffic Class Conversational Streaming Interactive Background
Characteristics Preserve time re- lation (variation) between informa- tion entities of the stream
Preserve time
relation (variation) between informa- tion entities of the stream Request/Response pattern Destination is not ex- pecting the data within a certain time Conversational pattern (stringent and low delay)
Preserve payload content
Preserve pay- load content
Mode Circuit Circuit, Packet Packet Packet
Type of
asymmetry
Symmetric Highly asymmetric Asymmetric Highly asym-
metric Sensitivity to
errors
Slightly sensitive Sensitive Very sensitive The most
sensitive Sensitivity to
delay
The most sensi- tive
Very sensitive sensitive The least sen-
sitive
Example Voice Streaming audio
and video
Web-browsing SMS, back-
ground download of emails
establish and manage several radio connections (or radio bearers) at the same time. For instance, a mobile user can, at the same time, download a file using ftp service and talk with another user via video teleconference.
The different types of UMTS services are grouped into four UMTS service classes according to their characteristics: conversational, streaming, interactive and background. This classification leads to an efficient management of radio resource units. The different classes take into account the restrictions and the limitations of the radio interface [TS204]. The main distinguishable factors among these classes is the sensitivity to delay, sensitivity to errors and asymmetry between uplink and downlink traffics (table 1.1).
The conversational class is the most sensitive class to delay, because it includes real-time services; the conversational nature of this class and the time relation between information entities of the stream must be preserved and thus, the delay must be low. Furthermore, the data processing at UMTS radio protocol stack should be as fast as possible. The most well known service of this class is the basic telephony service. Moreover, other multimedia and IP-based services such as voice over IP and video conferencing tools are also included in the conversational class. The maximum supported delay depends on the human perception of audio and video conversation. Besides, this service class is the least error sensitive class.
Another real-time service class is the streaming class. The QoS requirements of this class are also related to the human perception, since the destination of this class is always a human. In the streaming class, the time relation between information entities of the stream must be preserved to conserve the quality of the audio or the video. Therefore, a buffering process is used at the receiver side. Moreover, the streaming class requires sufficiently low bit-error-rate levels in order to offer satisfying QoS levels to users.
dropped. As we will see in section 2.4.2, bit error and block error rates are increasing function of the CIR. Therefore, the QoS levels of all service classes in the radio interface can be evaluated using the CIR.
Another important characteristic of UMTS services is the asymmetry between uplink and down- link traffics. The services of the conversational class are typically symmetric due to the conversa- tional nature of this class. Unlike the conversational class, other service classes require asymmetric traffic, with very high downlink to uplink load ratio; in some services, only signaling and request traffics are transmitted in uplink (e.g. Streaming services) (table 1.1). Obviously, spectrum ef- ficiency is very sensitive to the asymmetry between the two link directions. Therefore, a fixed symmetric allocation for uplink and downlink bandwidths (e.g. UMTS FDD mode) may lead to a saturation of the loaded link whereas the second link is almost unused. Hence, systems that offer higher flexibility in resource distribution between uplink and downlink are more suited for asymmetric services (e.g. UMTS TDD mode). For this reason, TDD mode characteristics have been a hot subject in the last few years.
1.1.2 UMTS Network
UMTS network is divided into three interacting domains with standardized interfaces. The stan- dardized interfaces facilitate network maintenance and allow operators to have different suppliers. The three interacting domains are: User Equipment (UE), UMTS Terrestrial Radio Access Network (UTRAN) and Core Network (CN) as depicted in figure1.2[HT00][LWN02]. UTRAN provides air interface access procedures for UE. CN insures information routing and switching between different UTRANs and between UTRAN and other telecommunications and service networks (GSM, fixed network, Internet network, Intelligent Network, etc.). CN also includes databases and network management functions. UEs and UTRAN are connected via the Uu interface, while UTRAN and CN are connected via the Iu interface. In this dissertation, only UTRAN and the Uu interface are investigated.
UTRAN is composed of several Radio Network Sub-systems (RNS) which can be interconnected via the Iur interface. This interconnection allows CN independent procedures between different RNS, such as handover. Therefore, CN is transparent for radio access technology-specific functions [HKK+00]. Each RNS comprises only one Radio Network Controller (RNC) and one or more base
stations, which are referred as Node Bs.
A node B performs functions relevant to physical layer processing, such as spreading and mod- ulation, coding and interleaving, rate adaptation, physical measurements and error handling. This network entity also performs some basic radio resource management operations such as micro- diversity/softer handover or closed-loop power control. Each Node B can serve one cell or more and may support either the TDD mode or the FDD mode, or both modes simultaneously.
1.1 Universal Mobile Telecommunications System 5 Core network NodeB Iub Iu UTRAN Uu RNC RNS Iur RNC External networks UE UE UE
Figure 1.2: UMTS architecture
macro-diversity/soft-handover, ciphering, broadcast signaling, open-loop power control, etc. Node Bs of a given RNS are connected to the corresponding RNC via the Iub interface.
UTRAN is designed to support both FDD and TDD modes on the radio interface. Both modes use the same network architecture and the same protocols. Only the physical layer and the air interface Uu are specified separately.
1.1.3 Radio Interface Protocol Architecture
The radio interface protocol stack includes three layers as depicted is figure 1.3 [TS202h]: the physical layer (layer 1), the data link layer (layer 2) and the network layer (layer 3).
Physical Layer
All physical layer procedures are handled by Node Bs. The physical layer [TS202f][TS202g] offers services to the above MAC layer via radio transport channels, and controls the transfer of physical channels over the radio interface. The physical unit used in the air interface is a 10 ms frame. The period of a transport channel is a multiple of the basic radio frame period and is called Transmission Time Interval (TTI), where TTI ∈ {10, 20, 40, 80} ms. Small values of TTI are suited for real time services such as voice in order to reduce the maximum delay.
Data Link Layer
The data link layer is split into four sub-layers: Radio Link Control (RLC), Medium Access Control (MAC), Packet Data Convergence Protocol (PDCP) and Broadcast/Multicast Control (BMC). Mac Sub-layer: The MAC sub-layer [TS202a] is responsible for mapping logical channels onto appropriate transport channels. Logical channels are classified into control and traffic channels. The logical control channels transport Control-plane (C-plane) information, whereas the logical traffic channels transport User-plane (U-plane) information. One of MAC functionalities is priority handling between different data flows that share the same physical channel. The priority handling
Figure 1.3: Radio interface protocol architecture
procedure is based on physical constraints, which are communicated via physical measurements and the QoS requirement fixed by higher layers. Another functionality of the MAC layer is the multi- plexing and the demultiplexing of logical channels into and from the transport channels delivered to and from the physical layer.
RLC Sub-layer: The RLC sub-layer [TS202b] provides segmentation/reassembly and retrans- mission services for both user and control data. The radio resource control layer configures RLC instances in one of three modes: transparent mode, unacknowledged mode and acknowledged mode. In the transparent mode, no protocol overhead is added to higher layer data. In the unacknowledged mode, retransmission protocols are not used and error free data delivery is not guaranteed. Finally, the Automatic Repeat reQuest (ARQ) mechanism is used in the acknowledged mode in order to decrease the error rate. For all RLC modes, the physical layer estimates the Cyclic Redundancy Check (CRC) error detection and the result CRC is delivered to the RLC together with actual data [LWN02].
PDCP Sub-layer: The PDCP sub-layer exists only in the U-plane. The main function of this sub-layer [TS202c] is the compression of redundant protocol control information (e.g. TCP/IP and RTP/UDP/IP headers) in the transmission entity and the decompression in the receiving entity. Header compression decreases the header load and thus allows utilizing scarce radio resources more efficiently when transmitting IP packets over the radio interface. PDCP compression is possible because few TCP/IP header fields changing from one IP packet to another and the rest of header fields remain more or less the same.
BMC Sub-layer: The BMC sub-layer [TS202d] controls cell broadcast center messages over the radio interface and handles only U-plane information.
Network Layer
The network layer only contains the Radio Resource Control (RRC) layer. The RRC layer [TS202e] is the ”head” layer in the UTRAN stack and controls all other layers. The main function of the