Dynamic Bandwidth Allocation of AAL2 connections in the UTRAN

Document Sample
Dynamic Bandwidth Allocation of AAL2 connections in the UTRAN Powered By Docstoc
					      Dynamic Bandwidth Allocation of AAL2 connections in
                        the UTRAN
                                                Stephen Waller, Neco Ventura
                                    University of Cape Town, Dept. Electrical Engineering
                                        Rondebosch 7700, Cape Town, South Africa
                                              {swaller, neco}
                                          Tel. (021) 650 2813, fax. (021) 686 1730
                                                                           Bandwidth allocation could be performed according to the
   Abstract— With the advent of 3G networks (UMTS,                      peak cell rate of the VBR sources. Such an approach leads
the Universal Mobile Telecommunications System)                         to underutilization of wireless resources due to the bursty
major players in the telecommunications industry have                   nature of the sources. The wireless bandwidth will be wasted
envisioned a system that is far ahead of current                        and the wireless network will experience high call blocking
technology in areas such as efficiency, intelligence and                and forced termination probabilities. Bandwidth allocation
power, with large amounts of bandwidth and more.                        could be performed based on the sources mean cell rates. In
Although this may be true in some aspects, in reality it is             such an approach, certain types of applications e.g. video
just another technology on the telecoms route to the
                                                                        sources will suffer from unacceptable losses and delays
future and problems still exist. Here we specifically
                                                                        (especially those with hard real-time constraints). This paper
focus on bandwidth, it being limited and costly.
                                                                        however, addresses the bandwidth allocation issue for the
   In order to manage bandwidth in telecommunications
                                                                        fixed part of the access network.        We propose a DBA
networks many techniques have been investigated. This
paper explores the applications of A2S (AAL2 signalling)                scheme that provides efficient usage of the transmission
as a tool for DBA (dynamic bandwidth allocation) within                 network by reallocating the bandwidth that was allocated to
the UTRAN. SHO (Soft Handover) is identified as a                       the transport network connections during the establishment
good time to renegotiate and reallocate bandwidth, as                   phase. This scheme also provides high wireless network
this is when new connections are being established. The                 utilization by ensuring that the delay requirements of the
ultimate goal is to achieve more efficient bandwidth                    transport network are not violated.
utilization.                                                               The paper is organized as follows. Section II covers the
                                                                        UTRAN (UMTS Terrestrial Radio Access Network) and
  Index Terms— UTRAN, ATM, AAL2, A2S, Delay,                            makes note of its performance requirements. Section III
Dynamic Bandwidth Allocation, SHO.                                      focuses on handover in the context of WCDMA as well as
                                                                        the issue of delay and how it relates to bandwidth efficiency.
                                                                        Their relationship with DBA is explained. The focus then
                     I. INTRODUCTION                                    shifts to AAL2 and AAL2 switching and signalling where
                                                                        the technology and its applications are presented in section
I  n a wireless network, bandwidth is perhaps the most
   precious and limited resource of the whole communication
system. Therefore, it is of extreme importance to use this
                                                                        IV. Section V explores the usage of AAL2 as a tool for DBA
                                                                        within the UTRAN, while section VI presents an
resource in an efficient way. As most of the current                    experimental framework. Finally Conclusions and future
applications produce Variable Bit Rate (VBR) traffic with a             directions are presented.
considerable degree of burstiness and also the AI (Air
Interface) capacity varies in time we find that the transport
network carries traffic of an extremely variable nature [20].                     II. THE ACCESS NETWORK FOR UMTS
This, together with user mobility and usage patterns makes                 UMTS (Universal Mobile Telecommunications Network)
the mobile network a highly dynamic environment.                        is the 3rd generation of cellular networks designed to be a
   To guarantee Quality of Service (QoS) in such an                     stepping-stone in the convergence of networks of all media
environment is no trivial matter and resources should be                types. Integration/convergence is necessitated, for one, by
managed and allocated in such a way that allows them to be              applications that support both voice and data type services.
used efficiently while meeting delay requirements. In this              Benefits of convergence include a reduction in toll charges,
paper we focus on resource (specifically bandwidth)                     cable plant costs and network support staff requirements
allocation as a solution to providing this QoS and making               [24]. UMTS promises to have higher data rates, with similar
more efficient usage of bandwidth resources.                            QoS (Quality of Service), than existing networks (cellular
                                                                        and landline). Voice and other circuit switched services, as
                                                                        well as packet switched data are supported. It addresses the
The authors would like to thank Telkom SA, Siemens, National Research   need for efficient use of the scarce radio spectrum and the
Foundation (NRF) and The Department of Trade and Industry (DTI) for
supporting this research project.                                       need for higher bandwidth, in addition to the need for a
                                                                        platform that efficiently supports data services. DBA can be
used in conjunction with CAC (Connection Admission               moment, this can be described as HHO. HHO can be
Control) and RRC (Radio Resource Control) to provide this        performed between two intra (same) frequency or inter
promised QoS by ensuring that bandwidth, delay and jitter        (different) frequency cells, and between two different
requirements are met. More efficient use of bandwidth can        systems (e.g. UMTS and GSM). This is the type of
also be achieved.                                                handover most frequently encountered in wireless systems
   The UTRAN (UMTS Terrestrial Radio Access Network)             due to its relative simplicity and because SHO is usually
is the access network specific to UMTS. Figure 1 shows a         only employed within CMDA (Code Division Multiple
schematic view of the UMTS network with particular focus         Access) systems. HHO is usually necessary in cases when
on the access part. It consists of nodes and the wire-line       1) the network is loaded and it is required to switch to a
interfaces between these nodes. The UTRAN interfaces             different carrier frequency, 2) the user is moving too fast to
with the core network using the Iu (in UMTS terminology)         be supported on a specific tier, 3) the coverage area of a
interface and the mobile device via the WCDMA AI                 particular frequency is exited and 4) the QoS provided by a
(Wideband Code Division Multiple Access Air Interface,           cell is not sufficient or appropriate. It is clear that although
also know as the Uu interface). The Iu interface consists of     the frequency reuse in a CDMA system is supposed to be
both a packet-switched and a circuit-switched section.           one, this is not always practical or possible, thus
                                                                 necessitating HHO. The drawback of HHO (when compared
                                                                 to SHO) is that the mobile is disconnected from the network
                                                                 for a definite period of time. This increases the chance of
                                                                 the call being dropped.
                                                                    Soft handover (SHO), on the other hand, is when the
                                                                 process of handover is carried out and the MT is connected
                                                                 to several Node Bs simultaneously. These Node Bs can be
                                                                 under the control of the same RNC or under the control of
                                                                 different RNCs. In the latter case the first RNC is referred
                                                                 to as the SRNC (Serving RNC) while the second is referred
                                                                 to as the DRNC (Drift RNC). SHO is possible between intra
                                                                 frequency cells [2] and reduces the chances of the call being
                                                                 dropped (handover failure) and also data loss during the
                                                                 handover period [3]. SHO is enabled by a technique known
                                                                 as macrodiversity (when referring to CDMA systems) [7].
                                                                 This macrodiversity (together with power control) comes
   The function of the UTRAN is to connect mobile users to       with added benefits of increased capacity and coverage, thus
the UMTS core network and support mobility (location             making efficient use of the scarce radio spectrum and
management and handover). It must do this while making           keeping inline with the ideals/goals of UMTS. In fact,
efficient use of resources (such as the bandwidth on the         “SHO is necessary for satisfactory operation in a fading
ATM (Asynchronous Transfer Mode) links between the               environment”. It allows the received signal strength and
nodes). A basic principal in the design of UMTS is the           quality to be raised to a level not possible or acceptable
separation of radio issues from the core network. This           without SHO. The main disadvantage of SHO is its need for
allows different access network technologies to be used with     extra channels [14][12].
one core network [15].                                              The amount of time spent in SHO varies depending upon
                                                                 the network layout (radio planning) and the users mobility
                                                                 pattern. Thus if cells overlap a lot and/or a user remains on
  III. HANDOVER AND DELAY IN RELATION TO DYNAMIC                 the outskirts/edge of a cell it is possible that a MT be in
              BANDWIDTH ALLOCATION                               SHO for the entire duration of the call. At this extreme it
                                                                 has been proposed to use macrodiversity for the entire
  A. Handover in WCDMA                                           duration of a call in order to improve coverage and link
   In this section we consider handover because it is during     quality. In order to do this the coverage area of cells in the
handover that new connections are established and old            access network are designed to overlap at every spatial point
connections released. In order to reduce the signalling          [3]. Also, in UMTS, 30% to 50% of all calls may be in
requirements for DBA it is sensible to incorporate it into the   SHO at a particular time [14]. From this we can conclude
process of SHO. Within a WCDMA (Wideband Code                    that the duration of time spent in SHO can be substantial. It
Division Multiple Access) system, widely accepted (and           is clear that SHO cannot be seen as simply another type of
specified) as the AI in the UTRAN, we encounter several          handover, but rather a central function of the radio access
types of handover. The types of handover within WCDMA            network (part of Radio Resource Management (RRM)) that
can be classified into two groups: SHO (soft handover) and       allows more efficient use of the radio spectrum as well as
HHO (hard handover). A point to note here is that SHO can        better quality of service for the end user. It is vital to the
only occur in WCDMA FDD (Frequency Division Duplex)              correct and efficient functioning of the access network. For
mode, not TDD.                                                   this reason this paper will focus on this type of handover in
   When a handover occurs and the MT (Mobile Terminal)           the discussion that follows.
is connected to only one Node B (3G base station) at any
  B. Bandwidth Efficiency versus Delay                             mitigation.
   One of the major performance issues regarding the
transmission requirements of the transport network in the
UTRAN is the tradeoff between delay and efficient use of            IV. AAL2; THE PROTOCOL, SWITCHING, SIGNALLING AND
bandwidth. It may seem that meeting the tight delay                                   APPLICATIONS
requirements is all that needs to be accomplished, but in-            Much research work has been done in the area of the
fact, without sufficiently relaxing the requirements on delay,     AAL2 protocol and many papers refer to its efficient use of
bandwidth is utilized poorly. Using DBA we can distribute          bandwidth and ability to address the problem of
and use bandwidth efficiently and appropriately thus helping       packetization delay. It is clear that AAL2 is a good
avoid and alleviate congestion situations. This has a positive     candidate for carrying low and variable bit rate, real-time
effect on delay.                                                   (delay-sensitive) traffic, such as that encountered in cellular
   There are several reasons why there is such a strong delay      networks [8][9][10][16][17].
requirement within the UTRAN: First of all the required               At the core of AAL2 is its ability to multiplex variable
radio level synchronization imposes stiff delay and jitter         size packets from multiple users allowing more efficient use
requirements.        This requirement for radio level              of bandwidth resources. For these reasons the wire-line
synchronization is brought about through the SHO                   interfaces within the UTRAN have been specified to use
functionality and the use of macrodiversity combining [20].        AAL2 as one of the transport technologies for circuit
A point to consider in this case is the MDU (Macrodiversity        switched connections. Other applications where AAL2 has
Unit) situated within the RNC [3]. This device works in real       been applied is in ATM Trunking and Voice over DSL
time and the combining of AI frames from different legs of a       (VoDSL) [8]. The main benefit of using AAL2 is that
SHO connection requires a certain bound on the transport           instead of just transporting a single packet from one user, the
delay across the Iub and Iur interfaces. The greater the           ATM payload can be filled by many users (thus using the
difference in the transport delay between two different SHO        bandwidth more efficiently through statistical multiplexing).
legs, the longer the MDU takes to do it’s job and ultimately       Another important benefit is the control of delay; depending
more of the delay budget is used [25]. Secondly we note            on the number of AAL2 connections using and ATM VC,
that radio frames arriving at the MT from different cells          one can expect good delay and delay jitter performance. In
should arrive synchronously in order to minimize UE (User          cases where less than optimal numbers of AAL2 connections
Equipment) buffers [22]. Thirdly, due to the complex               are present the use of a timer (Timer_CU in AAL2
scheduling function of the MAC layer (schedules based on           terminology) can ensure that the delay requirements of
traffic class, RF interference, power control efficiency,          packets are met.
synchronization, and others), the AI (Air Interface)                  Figure 2 is a diagram of the user plane protocol stack
resources are reserved ahead of time. Thus if frames are           between the UTRAN Nodes. It clearly shows the location of
excessively delayed (or lost) this reserved capacity is wasted     the AAL2 protocol within the UTRAN. ATM and AAL2
[19]. This sets an upper bound on the transmission time for        are not present on the AI between MT (Mobile Terminal)
the radio frames.                                                  and Node B (or BS). The AAL2 switch represents the
   From the above we can see that the delay requirements for       transport network between the UTRAN nodes and is not
the transport network in the UTRAN apply not only to RT            necessarily a single switch. It can be a combination of ATM
traffic, but also to NRT (best effort) traffic. Thus all traffic   and AAL2 switches as well as the switching functionality
must be treated as having RT QoS requirements                      within a D-RNC (Drift RNC).
   Although there is this strict delay requirement on the
UTRAN transport network there is a tradeoff that needs to
be made. Congestion in the network results in a high
bandwidth requirement and poor bandwidth utilization [21].
An excessively strict requirement on delay further
aggravates this congestion condition resulting in further
bandwidth needs (due to poor utilization of available
bandwidth). If the delay requirement is relaxed it is thus
possible to achieve higher bandwidth utilization during these
times of congestion. Ultimately we need enough bandwidth
to meet delay requirements but at the same time, in order to
achieve a higher utilization of bandwidth, this requirement
needs to be relaxed.
   From this discussion we can see that correct and efficient         In order to setup these AAL2 connections on demand
bandwidth utilization is extremely important in order to           (dynamically) as opposed to statically, A2S (AAL2
avoid congestion. We also note that avoiding congestion            signalling) is required. Without A2S the AAL2 connections
situations as far as possible is beneficial in avoiding            would have to be statically provisioned ahead of time
excessive delay in the UTRAN. We aim to show that DBA              (similar to a PVC (Permanent Virtual Circuit)), thus wasting
can be used as a tool in congestion avoidance and                  ATM bandwidth resources while not in use. Also the
                                                                   bandwidth assigned to the AAL2 connections is fixed and
cannot be changed. In simple terms, A2S manages the               bandwidth, (c) The dynamic nature of the AI imposes further
allocation of AAL2 connections to incoming requests               requirements/constraints on the user application and data in
(similar to SVCs (Switched Virtual Circuits)). The on-            terms of flexibility with regard to available resources. What
demand allocation of these AAL2 connections allows the            flows over the AI flows over the Iur and Iub and thus the
efficient and flexible use of ATM resources. Another              traffic is dynamic in this regard also.
important point to note about A2S is that it allows a fast
                                                                    C. A2S and DBA in the UTRAN
connection setup and release as required in environments
such as the UTRAN where the requirements of SHO need to              In order for bandwidth to be flexible it makes sense to be
be taken into account. AAL2 switching is required most            able to change the bandwidth allocated to connections
importantly as a bandwidth saver [8][9].                          dynamically. Since AAL2 (User plane) is used as the
   It is clear that AAL2 is an appropriate tool for use within    transport network within the UTRAN and A2S (Transport
the UTRAN. With this established we now look further at           Network Control plane [27]) is used to establish, manage
A2S and its ability to support DBA [26]. When used in             and teardown these connections (transport bearers) it would
conjunction with SHO, we would like to show that DBA              be reasonable to use A2S in order to dynamically manage
allows further efficiency in the use of bandwidth and that it     the bandwidth allocated to these connections.
can help in avoiding and alleviating congestion.                     Within the process of DBA, A2S modifies (increasing or
                                                                  decreasing) the bandwidth allocated to an AAL2 connection
                                                                  (AAL2 connection resources) during the established phase.
     V. DYNAMIC BANDWIDTH ALLOCATION AND A2S                      This signalling process is carried out on request from the
                                                                  A2S served user (Radio Resource and Handover control
  A. Bandwidth is limited                                         entities within the UTRAN [20]). Our proposal is to
   Bandwidth is one on the most important resources in any        perform DBA, specifically during HO periods, using A2S as
network or telecommunications system. Within a cellular           a tool to achieve our end goal, which is efficient (or
network the bandwidth on the AI is extremely costly and           higher/better) utilization of network bandwidth.
limited and network designers (and others) are constantly            The following research phases will be investigated:
striving to find ways to make better use of this scarce           1) Emulate congestion and implement a suitable algorithm
resource. This section explores the role of A2S as a tool for         that uses the DBA properties of A2S. Thus determine
performing DBA on the fixed links (specifically Iur) within           the usefulness of A2S DBA in the context of congestion
the UTRAN. Ultimately, A2S together with Handover and                 avoidance and mitigation.
Radio Resource Control entities should allow an                   2) Determine in what cases (e.g. usage pattern, application
improvement in the use of bandwidth resources in terms of             behavior, network conditions) it would be possible (or
efficiency and congestion.                                            useful) to decrease/increase the bandwidth of a
   One should not be led to believe that, since the AI is             connection during SHO and how is this affected. This
considered as the bottleneck in a cellular system, the                point focuses on SHO and how DBA functions in this
bandwidth on the Iub and Iur interfaces within the UTRAN              context.
should be considered infinite or even abundant. Firstly,          3) Determine in what cases the bandwidth of a connection
there is always room for more bandwidth (in any system)               can be decreased or increased depending upon usage
and even if over-provisioning is adopted the excess                   pattern. How does this affect the users perceived QoS
bandwidth will be utilized as applications and traffic loads          and the network load? Is it possible to change the
demand more. Secondly, it is not good engineering practice            concept of a fixed traffic contract (adaptive QoS)?
to overprovision in order to satisfy a system’s requirements          Rather than what applications, what type (this approach
– resources should always be used wisely and efficiently.             is more general and useful) of applications can be used in
Finally, not only should this resource be used efficiently but        this scenario? There is a need to explore more about
also, as we have seen, the management of this resource has            adaptive traffic descriptors/profiles and contracts.
direct consequences on the AI. Remember also that the             4) There is a tradeoff between the signalling overhead
terrestrial land line network may be provided by another              required to implement DBA and the improvement in
network operator (possibly sharing them with other service            resource utilization as a result of using it. Thus it would
providers) and thus these leased lines may be expensive and           be useful to determine the signalling load (cost) due to
one needs to define the bandwidth required on these links in          the use of A2S to enable DBA within the AAL2
order to specify to the network provider the bandwidth                connections. In what cases it is not economically
required.                                                             feasible to use this DBA in order to save bandwidth in
                                                                      the user plane? In other words, at what point does the
  B. Bandwidth is dynamic                                             bandwidth being used in the control plane to implement
  Bandwidth management in a cellular environment should               A2S DBA equate to or exceed the bandwidth being
be flexible as this resource is extremely dynamic. It is              saved in the user plane? Increasing the rate of bandwidth
dynamic in several regards: (a) User applications demand              re-allocation and noting the required bandwidth (cost) in
varying amounts of bandwidth, (b) With many mobile users              relation to the bandwidth saving in the user plane will
not only is the bandwidth varying due to call arrival and             address this task.
departure rates, but also the mobility of the users adds to the      A DBA algorithm, situated in the control plane of the
dynamic nature of the system and its requirements on              RNC, would monitor conditions in the network and detect
situations in which DBA could be used. This algorithm,          can be used to make more efficient use of bandwidth. An
together with the radio resource and handover control           attempt will be made to quantify the costs and benefits from
entities (also residing within the Control plane) would then    this approach.
pass a request to A2S to modify connection resources under         The main point of this experimental framework is to show
its control.                                                    the application of A2S to DBA within the UTRAN and
    In the past establishing a new connection with the          ultimately its ability to achieve the end goal of better (more
required bandwidth, transferring the user from the old to the   efficient) bandwidth utilization. It will show how AAL2
new connection, and then releasing the old connection           DBA functions and how it could interact with higher layer
accomplished the process of DBA. This clearly uses a lot of     control functionality.
resources in terms of available bandwidth (two connections
up simultaneously). A2S Capability Set Two (CS2) allows
the modification of bandwidth within the same connection                             VII. CONCLUSIONS
(without setting up a completely new connection). This uses        We have talked about mobility within the UTRAN and the
less bandwidth and the signalling load is also reduced. As a    requirements it imposes on the transport network. With the
result resources (specifically bandwidth) are used more         high requirements on delay within the UTRAN the effects on
efficiently and more users can be accommodated with the         bandwidth utilization and congestion are often overlooked.
same bandwidth.                                                 In this paper the tradeoff between delay and efficient usage
                                                                of bandwidth has been highlighted. The advantages of
                                                                applying AAL2 and its associated switching and signalling
             VI. EXPERIMENTAL FRAMEWORK                         protocols within the UTRAN have been mentioned. Also,
   Based on the foregoing discussion, we propose an             the application of A2S as a tool for DBA, specifically for
experimental framework, supporting AAL2 and the A2S             congestion avoidance and more efficient usage of bandwidth
protocols. Functionality to support SHO and AAL2 DBA is         has been introduced.
included. Scalability and complexity will be taken into            It is concluded that AAL2 is indeed an appropriate
account.                                                        technology for use within the UTRAN. Together with
   The minimum requirement in order to show SHO within          AAL2 switching it provides an efficient means of
the UTRAN is one RNC and two Node Bs. Thus our                  transporting mobile network traffic. A2S allows flexible
experimental framework consists of four nodes being             control of AAL2 connections, necessary in order to utilize
connected via ATM links. This is shown in figure 4 below.       statistical multiplexing on the ATM VC level and thus
Each node has A2S functionality and the ATM links carry         allowing more efficient bandwidth utilization. The ability of
AAL2 paths. The different nodes carry only the necessary        A2S to support DBA allows further improvements in
functionality to perform the required research.                 bandwidth utilization and can be used as a tool for
                                                                congestion avoidance and mitigation.
                                                                   SHO has been identified as a good time for the proposed
                                                                protocol to be triggered. Further research needs to be done
                                                                into adaptive traffic descriptors/profiles and traffic contracts
                                                                in order to determine suitable applications for use with A2S

                                                                [1] Bo Karlander, Szilveszter Nadas, Sandor Racz and
                                                                    Jonas Reinius, “AAL2 switching in the WCDMA radio
                                                                    access network”, Ericsson Review No. 3, 2002
                                                                [2] Kimmo Hiltunen, “Handover in WCDMA Uplink and
                                                                    Downlink. Performance of Handover Algorithms”,
                                                                    Postgraduate Course in Radio Communications, 2001 –
                                                                [3] Ulrich Weiss, “Designing Macroscopic Diversity
                                                                    Cellular Systems”, VTC’99, Houston, Texas, USA,
                                                                    May 16-20, 1999
  A traffic source will be located within the core network      [4] N. Efthymiou, Y.F. Hu, A. Properzi, E. Buracchini, S.
node, which will generate traffic to be sent via the RNC to a       Faccin, M.T. Scarpelli, P. Zanini, W. Zhao, R.
specific (or both, during SHO) Node B. Functionality                Tafazolli, “Inter-Segment Handover Implementation in
within the RNC includes AAL2 switching, Handover and                the SINUS Project and its Integration within the
Resource control modules and the DBA algorithm. The                 RAINBOW Project” 1997, available online:
DBA algorithm will interact with Handover and Resource    
control modules. Within the process of SHO, A2S will then           .PDF
be called upon to establish and manage the relevant AAL2        [5] Ian F. Akyildiz, Janise Mcnair, Joseph S. M. Ho,
connections. This SHO process will be examined and                  Huseyin Uzunalioglu, Wenye Wang, “Mobility
specifically how A2S functions in this context and how it
     Management in Next-Generation Wireless Systems”,            [24] David Willis, “Integrating Voice and Data”, October
     Proceedings of the IEEE, Vol. 87, No. 8, August 1999             1999,                 available              online:
[6] 3G (IMT 2000, UMTS) Hierarchical Wireless Systems,      
     available online:          s=NCJS_1021bt
[7] Andrew J. Viterbi, Audrey M. Viterbi, Klein S.               [25] 3GPP TS 25.932 Delay Budget within the Access
     Gilhousen, Ephraim Zehavi, “Soft handover extends                Stratum.
     CDMA cell coverage and increases reverse link               [26] ITU-T, “AAL Type 2 Signalling Protocol – Capability
     capacity”, IEEEE JSAC, vol. 12, No. 8, October 1994              Set 2”, ITU-T Recommendation Q.2630.2, December
[8] Andre van Zyl, “AAL2 Signalling Framework to                      2000
     Support a Gigabit AAL2 Switching Node”, MSc Thesis          [27] 3GPP TS 25.401 UTRAN Overall Description.
     University of Cape Town, Dec 2002
[9] Sven Shepstone, “AAL2 Switching Node to Support
     Voice Services in 3rd and 4th Generation Networks”,         Stephen Waller. Born in Natal, 03/09/1976. Completed the
     MSc Thesis University of Cape Town, 2002                    National Diploma in Electrical Engineering at the Port
[10] Stephen Waller, Xolani Fomana, “AAL2 Signalling to          Elizabeth Technikon in 1997 and the BSc Degree in
     support UMTS Terrestrial Radio Access Networks”,            Electrical Engineering at University of Cape Town in 2001.
     SATNAC Conference, September 2002                             He is currently studying toward the Masters Degree in
[11] Aku Poysti, “Blocking in Third Generation Radio             Electrical Engineering at the University of Cape Town.
     Access Networks”, MSc Thesis Helsinki University of
     Technology, June 2000
[12] Presentation by Kari Heiska, “ Radio Resource
     Management (RRM) Handover Control (HC)”, Nokia,
     March 2002
[13] Katerina Koutsaftiki, Paul Reale, “Planning for Profit”,
     3G strategies for operators, Lucent Technologies, q1
[14] Basics of UMTS, University of Melbourne, available
[15] Ericsson, “Basic Concepts of WCDMA Radio Access
     Network” White Paper, September 2001
[16] Seong-Ho Jeong, Deh-phone K. Hsing, Tsong-Ho Wu,
     John A. Copeland, “Trunking Performance Analysis of
     AAL2 for VBR Voice Transport”, available online:
[17] Notker Gerlich, Michael Menth, “The Performance of
     AAL2 Carrying CDMA Voice Traffic”, University of
     Wurzburg, Institute of Computer Science, Research
     Report Series, Report No. 199, April 1998
[18] Sumit Kasera, “Voice over AAL2 – With Special Focus
     on       AAL2       Signaling”,      available    online:
[19] Michael Menth, Matthias Schmid, Herbert Heib,
     Thomas Reim, “MEDF – A Simple Scheduling
     Algorithm for Two Real-Time Transport Service
     Classes with Application in the UTRAN”,
[20] G. Eneroth, G. Fodor, G. Leijonhufvud, A. Racz, I.
     Szabo, “Applying ATM/AAL2 as a Switching
     Technology in Third-Generation Mobile Access
     Networks”, IEEE Communications Magazine, Vol. 37,
     June 1999
[21] Soracha Nananukul, Yile Guo, Maunu Holma, Sami
     Kekki, “Some Issues in Performance and Design of the
     ATM/AAL2 Transport in the UTRAN”, IEEE Wireless
     Communications        and     Networking      Conference,
     September 2000
[22] 3GPP TS 25.402 Synchronisation in UTRAN
[23] Gabor Fodor, Gosta Leijonhufvud, Szabolcs
     Malomsoky, Andras Racz, “Comparison of CAC
     Algorithms in ATM/AAL2 Based 3rd Gen Mobile
     Access Networks”, IEEE, 1998