chapter 2 literature survey...
TRANSCRIPT
41
CHAPTER 2
LITERATURE SURVEY
2.1INTRODUCTION
In a mobile network, position of MNs has been changing due to
dynamic nature. The dynamic movements of MNs are tracked regularly
by MM. To meet the QoS in mobile networks, the various issues
considered such as MM, handoff methods, call dropping, call blocking
methods, network throughput, routing overhead and PDR are discussed
in this chapter.
2.2MOBILITY MANAGEMENT
IP mobility management protocols proposed by Alnasouri et al
(2007), Dell'Uomo and Scarrone (2002) and He and Cheng (2011)are
compared in terms of handoff latency and packet loss during HM.
Mobile IP is the current standard for supporting mobility in IP networks.
But it lacks support for fast handoff control, real-time location tracking
as per the investigation by Sami (1995) and Madhow (1994), and QoS
which are critical to real-time services in cellular networks. Stevens-
Navarro (2010) demonstrate that the Mobile IP also has high control
overhead due to frequent notification to the Home Agent and lacks
support for soft handoff that requires the ability to multicast traffic to
multiple base stations during handoff(in current and 3G cellular
networks, soft handoff is done on link-layer).
Hence, a good IP mobility management solution should have
fast location update, low signaling load when handoff takes place in the
42
network, IP QoS mechanism support and soft handoff support. The Fast
Hierarchical Mobile IPv6 (FHMIPv6) outperforms other protocols in the
case of packet loss and handoff latency. Effect of call arrival rate and
average queuing delay also explained and handoff latency and packet
loss had been analyzed for various protocols, such as Mobile IPv6
(MIPv6), Hierarchical Mobile IPv6 (HMIPv6), Fast MIPv6 (FMIPv6),
Fast Hierarchical Mobile IPv6 (FHMIPv6).The authors showed that
FHMIPv6 had minimum handoff latency in wireless networks. Efficient
buffer management and optimized forwarding time were needed for
minimizing packet loss and handoff latency during handoff management.
2.2.1 Mobile IP for Micro-mobility management
Mobile IP is the current standard for supporting mobility in IP
networks. But it lacks support for fast handoff control, real-time location
tracking, and QoS which are critical to real-time services and found in
cellular networks. Mobile IP also has high control overhead due to
frequent notification to the Home Agent and lacks support for soft
handoff that requires the ability to multicast traffic to multiple base
stations during handoff(in current and 3G cellular networks, soft handoff
is done on link-layer). Hence, a good IP mobility management solution
should have
Fast location update and handoff
Low signalling load on the network
IP QoS mechanism support
Soft handoff support
43
Based on the observation that most user mobility is local to an
administrative domain, hierarchical mobility management schemes were
proposed recently by Akyildiz and Wang (2002) and Sungwon et al
(2012),in which Mobile IP is used to manage the occasional macro-
mobility (inter-domain mobility) and cellular-network optimized
schemes were used to manage frequent micro-mobility (intra-domain
mobility). By localizing the scope of most location update messages
caused by micro-mobility within a mobile domain, these schemes
supported fast handoff and real-time tracking within a domain.
However, they do not have good support for IP QoS, fast
reservation restoration and soft handoff. In this paper, authors proposed a
connection tree based IP micro-mobility management scheme for mobile
users with QoS requirements.
2.2.2 Connection Tree concept for micro-mobility management
Qiang and Acampora (2002) reported to make easy fast
handoff and real-time location tracking, and Anantharam et al (1994)
reported to reduce the location update overhead in IP based mobile
networks, hierarchical IP mobility management schemes were needed.
The Hierarchical IP mobility management schemes described the Mobile
IP for macro-mobility (inter-mobile domains) and cellular-network
optimized schemes for micro-mobility within mobile domains. In this
work, authors applied the connection tree concept for wireless
Asynchronous Transfer Mode (ATM) networks to IP micro-mobility
management for mobile users with QoS requirements. The IP micro-
mobility management supports soft handoff, due to this (soft handoff
requires a mobile node to hold more than one channel at a time)
44
shortages of channels may occur. In this paper, authors concluded that
their schemes enabled fast and smooth handoff by using fixed Care of
Address (CoA) to localize most of the location update related operation
and pre-establishing connections to/from multiple base stations.
2.2.3 GSM Mobile Application Protocol (GMAP)
MM means network functions that enable users to roam in
different networks so that they can access telecommunications services
outside their home network. Roaming requires standardized ways of
transferring subscriber information between network registers such as
the HLR and the VLR. In the GSM 900 standard, and in its derivatives
DCS 1800 and PCS 1900, subscriber data related signalling is handled
by the Mobile Application Part (MAP) protocol described by Akyildiz et
al (1999).
Verkama et al (2002) discussed the requirements for the
mobility protocol of the third generation mobile network. The mobility
protocol was used to transfer subscriber data between network elements
and it forms the basis for roaming. A modular network design was
proposed where mobility is managed independently of the access and
backbone networks. The GSM Mobile Application Protocol (GSM
MAP) was found to provide a firm basis for evolution towards such a
platform. In GSM, MAP was used in one particular handoff, namely the
handoff between two Mobile Switching Centers. New GSM features
involving development of MAP were described and directions for further
MAP development were identified. The intelligent network was
visualized to be a corresponding platform for providing operator specific
services.
45
In this paper, authors examined the requirements imposed on
MM in Universal Mobile Telecommunications System (UMTS) and
evaluate the GSMMAP protocol from this perspective. GSM MAP is an
evolving protocol as new features are introduced in the ongoing GSM
Phase 2+ standardization.
2.2.3.1 GSM MAP as Mobility Management Platform
The existing mobile networks have mobility protocols for
mobility management. These protocols depend on the other system
components to a varying degree. The interfaces specified in the GSM
standard where GSM MAP applies are depicted in Figure2.1.The B-
interface between VLR and its associated MSC is internal.
Figure 2.1 GSM network interfaces where MAP applies
EIR Equipment Identity Register, HLR Home Location Register, MSC
Mobile Switching Center, VLR Visitor Location Register
The register structure shown in Figure2.1 is expected to satisfy
the needs of UMTS with two possible exceptions. The first one concerns
packet services for which new registers may be needed. The other
exception is the Service Control Point (SCP) of the Intelligent Network
(IN) concept, used for providing value added services. Both exceptions
46
are being addressed in GSM Phase 2+ standardization. Once they are
solved, the GSM MAP protocol provides adaptable platform for mobility
management as far as signaling needs between networks registers are
concerned.
In GSM, MAP is used in one particular handoff, namely the
handoff between two MSCs. This is the only part of MAP that is vaguely
related to the GSM air interface: currently it assumes a backward style
handoff with no macro-diversity. This kind of handoff does not cover all
possible radio access techniques. MAP/E can be extended to a general
handoff protocol to make MAP completely independent of the air
interface. In conclusion, MAP is a sound basis for evolution towards a
global UMTS mobility protocol. In particular, MAP is a well tested
mobility platform integrating already now over a hundred networks. Use
of MAP in UMTS would maintain compatibility with the GSM-based
systems. This scenario looks even more attractive if interworking
solutions between GSM and other standards emerge by the time of the
introduction of UMTS services. Naturally, MAP in its present form does
not fulfill all UMTS requirements, but it must be further enhanced.
2.2.4 IP Micro Mobility Management
Hui and Zhengkun (2006) reported that extended use of Multi-
Protocol Label Switching (MPLS) in mobile networks, the idea was
applying IP micro-mobility management to MPLS-based wireless access
subnet. A new QoS-aware MPLS-based micro-mobility management
technology was proposed, where a Local Mobility Management Server
(LMMS) storing the information of roaming hosts and several Paging
Servers (PSs) were included, capable of providing paging functionality
47
and eliminating the network bottleneck. QoS was guaranteed by static
Label Switched Path (LSP) configuration and network reliability was
improved with LSP redundancy. Using LSP forwarding mechanism, the
QoS grade will remain unchanged before and after handoff. The updated
protocol messages were proposed and the signaling procedures of
binding update and handoff were elaborated. Compared with the existing
schemes, the proposed technology had shown its merits in its high
reliability and flexible expansibility.
2.2.5 Improved Micro Mobility Management
Hairong et al (2007) performed an extensive research work to
improve the performance of micro-mobility management schemes on
handling fast moving mobile hosts within a local administrative domain.
However, earlier work by Yi-Bing (1997) was focused on reducing the
location update cost and handoff latency, without considering any
integrated form of QoS. Jong-Tae and Seung-Man (2010) observed that
the advancement of wireless access technologies and emergence of
numerous multimedia applications over Internet, mobile users required
not only the seamless mobility support but certain level of QoS
guarantee. MPLS protocol had been developed as the solution to IP
quality of service, gigabit forwarding, network scaling, and traffic
engineering in Internet core network. Motivated by the advanced
characteristics of MPLS technology, in this work, authors proposed a
MPLS-enabled micro-mobility management protocol which aimed to
enhance the end-to-end QoS provisioning and traffic engineering
capabilities when dealing with intra-domain movement in all-IP future
generation wireless networks.
48
2.2.6 Tree Based Routing Protocol
Hoon and Phan (2007) proposed a Tree-Based Routing
Protocol (TBRP) with node mobility management technique for
integrated mobile ad hoc and infrastructure networks. A suite of network
management protocols proposed by Beongku et al (2011) and Yu Gao
and Huaxin Zeng, (2009) built and maintained network architecture in
the form of small-sized trees, each starting from a node that could
directly communicate with a gateway. Each node in a tree kept track of
information for its descendent nodes. A new node registered with a
foreign agent along a tree path without resorting to an inefficient
flooding. Authors also developed an efficient routing protocol that
exploits tree information where route discovery did not use a flooding,
either. They examined the proposed protocol for its applicability and got
a promising result, even with high network traffic.
2.2.7 MaISAM for fast mobility management
The main goal of future “All-IP” networks is the provision of
any type of service anytime and anywhere. Satisfying this goal requires
the facing of many challenges. One of the main challenges is how to
deliver high data rates multimedia services for mobile users at low cost
without degradation of the QoS. In order to satisfy this challenge, a new
solution, that supports fast mobility management simultaneously with
efficient and fast re-reservation of resources during and after the handoff,
is needed.
Alnasouri and Mitschele-Thiel (2008) observed and several
solutions had been proposed to couple QoS with MM in a way that could
satisfy the real-time requirements. The most of these solutions depended
49
on ReSource reserVation Protocol (RSVP) which suffered from many
drawbacks. "Next Step In Signalling" protocol suite (NSIS) presented a
new promised framework that could be used to signal for different QoS
models and covered the drawbacks of RSVP. This work proposed a new
solution named Mobility management aware Next Step In Signaling for
"All-IP" Mobile communication networks (MaISAM). MaISAM
supported a fast and a smooth handoff simultaneously with a fast
reservation of resources during and after the handoff. It integrated the
Mobile IP Fast Authentication (MIFA) protocol with QoS-NSIS
Signaling Layer Protocol (QoS-NSLP). This was achieved by adding a
new object to accommodate MIFA messages. The main advantages of
MaISAM were its simple network architecture without introducing new
entities and its fast handoffs and fast resources reservation.
2.2.8 Single physical layer User-data switching Platform
Architecture Network (SUPANET)
Next Generation Network (NGN) requires the convergence of
different Access Networks (ANs) which may be wireless access
network. These ANs will interconnect each other through the Core
Network (CN). The AN and the CN should be able to provide the QoS
required for the different traffic passing through them and MM required
for MN across different ANs. Therefore, the QoS guarantee and MM are
very important areas in NGN.
Yu and Huaxin (2008) reported that guarantee of QoS and MM
are important research areas in Next Generation Networks (NGN). Most
of the people had been taking a route of Mobile IP over MPLS (MIoM)
to meet such challenge. However, MIoM had a drawback considering
50
that MPLS itself did not have any direct control over transmission at the
physical layer, instead, it had to rely on existing data link layer. Based on
the novel network architecture called Single physical layer User-data
switching Platform Architecture (SUPA), this paper proposed a new
Differentiated Dynamic QoS Provisioning (DDQP) mechanism. It is
tentatively concluded that DDQP could offer better QoS assurance than
MPLS and mobility management in NGN with regard mobile model.
2.2.9 Micro-mobility management schemes
Langar et al (2009) observes that an efficient MM was one of
the major challenges for next-generation mobile systems. Indeed, a MN
within an access network might cause excessive signaling traffic and
service disruption due to frequent handoffs. These two signaling traffic
and service disruption effects minimized to support QoS requirements of
emerging multimedia applications. In this work, authors proposed a new
adaptive micro-mobility management scheme designed to track
efficiently the mobility of nodes. This scheme minimized handoff
latency and total signaling cost when ensuring the MNs QoS
requirements. The authors introduced the concept of residing area. In
view of that, the micro-mobility domain was divided into virtual residing
areas where the MNs limited its signaling exchanges within this local
region instead of communicating with the relatively far away root of the
domain at each handoff occurrence. A key distinguishing feature of their
solution was its adaptive nature since the virtual residing areas were
constructed according to the current network state and the QoS
constraints.
51
To evaluate the efficiency of their proposal adaptive Master
Residing Area (MRA), they compared their scheme with existing
schemes (i.e., FMIP, MIP-RR, PF, Mobile MPLS and MMPLS) using
both analytical and simulation approaches for the 2-D random walk
model as well as real mobility patterns. Two scenarios are considered. In
the first one, only one MN moves between neighboring subnets and
receives 64 Kb/s down link packets. In this case, there is no background
traffic in all visited subnets. That is, only one connection exists in each
subnet at any time.
In the second scenario, authors investigated the case of
multiple connections in each subnet, which is more likely to be the case
in real networks. Explicitly, a MN moving at a low speed (i.e. 1 m/s) is
assumed to be present at each subnet and receives 64 kb/s background
traffic. According to each mobility scheme, the average registration cost
per handoff, the link usage and the handoff performance are calculated.
All the simulation results given below were obtained with very narrow
97.5% confidence intervals by Jang and Byung (2000) and Langar et al
(2009). To get an insight into the accuracy of their results under the2-D
random walk mobility model and the computation gain achieved through
the analytical study.
52
Table 2.1 Average cost of Cu and LU for adaptive MRA:
Analytical and Simulation Results
Radius R 2 3 4 5 6 7 8 9 10
LU (Analytical) 1.5789 2.3274 3.1433 3.8765 4.6216 5.3863 6.1445 6.8842 7.6498
LU (Simu) 1.5802 2.3265 3.1419 3.877 4.6244 5.3864 6.1387 6.8904 7.6461
Cu(Analytical) 456.3963 449.2111 460.7756 481.2784 505.6283 530.6132 556.5164 583.3888 609.6725
Cu (Simu) 456.0715 449.0087 460.6031 481.5721 505.8367 531.0735 556.6528 583.3818 609.8191
Time (Analytical)(sec) 0.53 2.41 11.5 49.77 132.13 412.47 1516.11 2953.63 5400.75
Time (Simu) (sec) 158.06 359.22 804.5 1663.55 2502.53 2759.75 4145.11 12796.98 14040.14
Cu - Signaling of registration update when handoff occurs
LU - Link usage in the micro-mobility domain
R - Micro mobility domain radius in km
Table 2.1 shows the analytical and simulation results regarding
the link usage and registration updates costs in the adaptive MRA case
when Dmax = R and when using the 2-D random walk. In the table, R is
varied from 2 to 10. The computation time for both analytical and
simulation results are listed in Table 2.1. The authors considered three
values of the radius R in their simulations: R = 2, R = 5 and R = 10.These
values of R are representative of small, medium and large micro-mobility
domains. They could observe that the link usage cost with FMIP, Mobile
MPLS,MIP-RR and M-MPLS schemes are the same and insensitive to
Dmax since, in these cases, packets are delivered using the shortest path
from the Label Edge Router Gateway (LERG) node to the current
serving LER/FA. In their adaptive MRA case, the link usage cost was
reduced considerably, compared to the PF scheme, notably when Dmax
was large.
53
2.2.10 Analytical model for Mobility management
An analytical modeling framework for the performability
modeling and evaluation of mobile and wireless mobile systems
supporting mobility, failure and repair behavior and queue capacity are
described by Kirsal et al (2011).
Kirsal, Y. et al (2011) also explained about the performance of
ability modeling and evaluation of wireless and mobile systems. Mobile
network, because of complexity of the next generation wireless and
mobile systems, performance evaluation is essential to improve the
architecture according to the QoS requirements and performance
characteristics. Dong and Maode (2012) observes that the vertical
handoffs occur between different technologies. It is very important to
provide seamless access to various technologies for next generation
networks (mobility between cellular network and WLAN). The work
presented by Hernandez et al (2008) in “Critical Review of Analytical
Modelling Approaches for Performability Evaluation of the Handover
Phenomena in Mobile Communication Systems” proposes an evaluation
framework of mobile computing systems with different handoff
strategies using analytical modeling approaches. The mobility issues
were also considered in the proposed models. Well-known approximate
Markov reward model solution and the exact Spectral expansion solution
approaches were considered by authors. In this work call blocking
probability is measured with various speeds of mobile nodes.
54
Figure 2.2Call blocking probabilities results with various Mobile
Node speeds
Figure 2.2 compares that when mean arrival rate increases, the
CBP also increases. In this Figure 2.2, S is defined as the number of
available channel, M is queuing capacity, R is radius of the cell, is
defined as the total call arrival rate in the cell, Toc defined as an average
call holding time in cellular network.
2.2.11 Cross Layer based Mobility Management
Peng et al (2011) observed that fast development of wireless
network technologies, a variety of different wireless networks coexist in
the future. The authors were getting more focus on convergence of
heterogeneous networks. In the environment of heterogeneous wireless
networks, traditional MM had already become incapable to support
mobility in heterogeneous wireless networks. HM had been one of the
55
most challenging technologies in heterogeneous network environments.
Fast, efficient and accurate handoff could give users a different
experience of networks, the authors aimed to propose a cross layer based
MM protocol which could meet the needs of the users to ensure that
users at anywhere and anytime access to the best network.
2.2.12 Human movement based Mobility management
Linet al (2012) observed that the human movement behavior is an
important issue for prediction of vehicle traffic and spread of
transmittable issues. Xian et al (2008) studies, since mobile telecom
network could efficiently monitor the movement of mobile users, the
telecom's MM would be an ideal mechanism for studying human
movement issues. The problem abstracted as follows: What was the
probability that a person at a location, ‘A’ would move to location ‘B’
after ‘T’ hours. The answer to this problem could not be directly
obtained because commercial telecom networks did not exactly trace the
movement history of every mobile user. In this work, authors showed
how to use the standard outputs (handoff rates, call arrival rates, call
holding time and call traffic) measured in a mobile telecom network to
derive the answer for this problem.
2.3 LOCATION TRACKING
In Mobile communications service system, it is required to
locate the destination of MNs when an incoming call arrives. Author
proposed a simple location tracking scheme called the T-threshold
location cache scheme. In the scheme, a threshold T is used to determine
whether a cached location record is obsolete. When the incoming call
56
frequency changes, this scheme adaptively modifies the threshold to
yield the best performance. An analytical model is proposed to study the
T threshold scheme. The analysis indicated that the T-threshold scheme
effectively reduced the network traffic compared to the Interim Standard
41 (IS-41) scheme. The IS-41 is an Inter Cellular Network
Communication Protocol which was developed to support
interoperability between differing networks for mobile telephony. The
protocol provides functionalities like call delivery to and from the
mobile network, intersystem handoffs and roaming, short message
delivery, validation and authentication through an inter-system
messaging protocol.
Specifically, IS-41 functions provide operations and procedures to:
1. Detect the presence of a mobile subscriber in a visited system.
2. Authenticate a subscriber for service.
3. Allow access to subscribed services while roaming outside the
original service area.
4. Provide continuity of communication through the handoff process
between systems.
In a PCS system, the location of a called portable MN must be
identified before the connection could be established. Yi-Bing (1994)
proposes an algorithm to efficiently locate a portable node. Author
introduced a PCS network architecture proposed in the Bellcore PCS
Network and Operations Plan which shows in Figure. 2.3.
57
Figure2.3 PCS network architecture
HLR: Home Location Register, RSTP: Regional Signal Transfer Point,
LSTP: Local Signal Transfer Point, MSC: Mobile Switching Centre,
SSP: Service Switching Point, RA: Registration Area, VLR: Visitor
Location Register
This model assumes that the HLR resided in the Service
Control Point (SCP) which is connected to a Regional Signal Transfer
Point (RSTP).Through a connection network, RSTP connected to all
Local STP (LSTP) in the region, which performed message routing
translation and screening functions in the Signaling System 7 (SS7)
network.
The individual STP's illustrated in Figure2.3 actually
represented the mated-pair configuration. The MSC and a VLR are
collocated with a Service Switching Point (SSP) (in the future, the MSC
is likely to evolve to become an SSP). Authors assumed that every SSP
serves exactly one Registration Area (RA). An RA consists of one or
more radio port coverage areas or cells. In this paper, the term RA is
used to represent SSP.
58
A procedure to locate a portable MN p is specified in IS-41by
Yi-Bing (1994). This procedure is referred to as the IS-41 scheme, and is
described in Figure 2.4.
Figure2.4 IS-41 scheme to locate a portable
1. The incoming call to a portable p is routed via the mobile
identification number to the home MSC.
2. If p does not reside in the corresponding RA, then the HLR is
queried for routing information.
3. The HLR queries the VLR of the RA where p resides.
4. The VLR in turn queries the MSC to determine whether p is
capable of receiving the call. If so, the MSC returns a routable
address Temporary Local Directory Number (TLDN) to the VLR.
5. The VLR relays the routing address back to the originating RA via
the HLR.
In the IS-41 scheme, network messages were sent from an RA
to query HLR. The work indicated that the message traffic due to
locating portables is significant. Thus, it is natural to use local caches to
reduce the network traffic. The idea is to store the locations of frequently
59
accessed portables in a local database (i.e. cache) within an RA. When a
call arrives, the location of the called portable is identified in the cache
to avoid sending query messages in the network. The steps to locate
portable using cache information are described in Figure 2.5.
Figure 2.5 Steps to locate a portable using the cached information
1. An incoming call arrives at Registration Area R1. The cache
record indicates that the portable p resides in Registration Area
R2.
2. The VLR of R2 is queried, which in turn queries MSC to find the
routable address for p.
3. The VLR of R2 returns the address back to R1.
Due to the mobility of portables, the location information in the
cache may be obsolete. If the obsolete cache information is used to
locate a portable, then the procedure shown in Figure2.5 fails, and after
the failure, the IS-41 scheme must be used to find the correct location. In
such a case, the overhead of using the cache information to locate a
portable is higher than the IS-41 scheme. Thus, the cached information
should be used as a hint to locate a portable, and some strategy is
required to determine whether a record is valid.
60
2.4 LOCATION MANAGEMENT
Akyildiz et al (1999) identifies LM as a two-stage process in
GSM network. It enables the network to find the current attachment
point of the mobile user for call delivery, as depicted in Figure 2.6(a).
Figure 2.6(a) Location management operations Figure 2.6(b) Research Issues
The first stage is location registration (or location update).
Here the mobile terminal periodically notifies the network of its new
access point, allowing the network to authenticate the user and revise the
user location profile. This has been demonstrated by Akyildiz et al
(1999) and Jin and Shashi (2009).
The second stage is call delivery. Here the network is queried
for the user location profile and the current position of the mobile node is
found. Current techniques for location management involve database
architecture design and the transmission of signaling messages between
various components of a signaling network. This has been demonstrated
by Xiaoxin and Victor (1995) and Xie et al (1993). As the number of
mobile subscribers increase, improved schemes are needed to support all
subscribers. Other issues are security, dynamic database updates,
querying delays, paging methods and paging delays. Figure 2.6(b) shows
these research issues with their respective location management
operation.
61
2.4.1 Location Update and Mobile Paging
In mobile communication, current mobile networks partition
their coverage areas into a number of Location Area. Each Location
Area consists of a group of cells and each MN performs a location
update when it enters a Location Area. When an incoming call arrives,
the network locates the MN by simultaneously paging all cells within the
Location Area. There are a number of inefficiencies associated with this
location update and paging scheme.
Excessive location updates may be performed by MN that are
located around Location Area boundaries and are making
frequent movements back and forth between two location areas.
Requiring the network to survey all cells within the Location
Area each time a call arrives may result in excessive volume of
wireless broadcast traffic.
The mobility and call arrival patterns of MN vary, and it is
generally difficult to select a Location Area size that is optimal
for all users. A perfect location update and paging mechanism
should be able to adjust on a per-user basis.
In addition, the Location Area-based location update and
paging scheme is a static scheme as it cannot be adjusted based on the
parameters of a MN from time to time. Many recent efforts focus
primarily on dynamic location update mechanisms which perform
location update based on the mobility of the MN and the frequency of
incoming calls.
62
2.4.2 Location Update Schemes
The standard Location Area based location update method does
not allow adaptation to the mobility characteristics of the MN. The
following techniques allow dynamic selection of location update
parameters, resulting in lower cost.
Dynamic Location Area management: This scheme introduces
a method for calculating the optimal Location Area size given the
respective costs for location update. This has been described by
Akyildiz et al (1999) Guohong (2001). The authors consider a
mesh cell configuration with square-shaped cells. Each location
area consists of cells arranged in a square and the performance is
analysed on a per-user basis according to the mobility, call
arrival patterns and the cost parameters. For example, it is
assumed that there are two mobile nodes MN1 and MN2, which
have different mobility and call arrival patterns. The Figure 2.7
shows the location area for MN1 and MN2. For MN1, the length
of the cells l value is 2 in location area and for MN2 the length of
the cells l is 4 in location area. It is generally not easy to use
different location area sizes for different mobile node as the MN
must be able to identify the boundaries of location area which are
continuously changing. The implementation of this scheme is
complicated when cells are hexagonal shaped, or when irregular
cells are used.
63
Figure 2.7 Registration Area
Figure 2.8 Time-based location update scheme
2.4.3Three dynamic update schemes
Three location update schemes are examined.
Time Based Update: AMN1 performs location updates
periodically at a constant time interval. Figure 2.8 depicts the
path of a MN1 from location A to D. If a location update occurred
at location at time 0, subsequent location updates will occur at
constant time interval.
Movement Based Update: A MN performs a location update
whenever it completes a predefined number of movements across
cell boundaries (this number is referred to as the movement
64
threshold). Figure 2.9 depicts the same path as shown in Figure
2.8. Assuming a movement threshold of three is used; the MN
performs location updates at the third movement of locations and
is shown in Figure 2.9. The MN movements are carried out from
A to C.
Figure 2.9 Movement-based location update scheme
Distance Based: AMN performs a location update when its
distance from the cell where it performed the last location update
exceeds a predefined value (this distance value is referred to as
the distance threshold). Figure 2.10 depicts the same path as
shown in Figure 2.8. A location update is performed at location
where the distance of the MN from location exceeds the
threshold distance. The MN movements are carried out from A to
B.
65
Figure 2.10 Distance-based location update scheme
The performance of the above schemes based on a simplified one-
dimensional movement mode is evaluated by Akyildiz et al (1999). The
distance-based scheme implementation illustrated by Vicente and Jorge
(2002) incurs the highest overhead. For the time-based and the
movement-based schemes, the MN has to keep track of the time stamp
and the number of movements from the last location update. This can be
achieved simply by implementing a timer or movements counter at the
MN. The distance based scheme, however assumes that the MN has
knowledge of the distance relationship among all cells. The network
must be able to provide this information to each MN in an efficient
manner.
2.4.4 Distance-based location update scheme
A distance-based location update scheme is considered by
Akyildiz et al (1999). The authors introduce an iterative algorithm that
can generate the optimal threshold distance that results in the minimum
66
cost. When an incoming call arrives, cells are paged in a shortest
distance-first order such that cells closest to the cell where the last
location update occurred are polled first. The delay in locating a MN is
proportional to the distance travelled since the last location update.
Results demonstrated that, depending on the mobility and call arrival
parameters the optimal movement threshold varies generally. However,
the number of iterations required for this algorithm to converge-varies
depending on the mobility and call arrival parameters considered.
Determining the optimal threshold distance may require significant
computation at the MN.
2.4.5 Per-user location caching
Per-user location caching reduces the signaling load due to
accessing the database in order to locate the called MN by maintaining a
cache memory close to the Signal Transfer Point (STP). For the sake of
simplicity in per-user location caching there is an MSC per VLR. Each
STP corresponds to an MSC/VLR. When a call is delivered, the STP
serving the calling MN stores in the cache memory the VLR where the
called mobile node is roaming. When a new call arrives, the STP checks
the cache memory. If the VLR where the called MN is roaming is stored
in the cache memory, then the VLR is queried and if the called MN is
still in the VLR service area the call is delivered (cache hit). If this
operation fails (cache miss), the GSM-MAP or IS-41 classical procedure
is followed.
The rate of incoming calls per MN and the rate of VLR
changes are analyzed. The rate of VLR changes are used as a parameter
67
to evaluate the cache hit probability. The validity period of the cache
entries is based on the mobility and incoming call rates. Maintaining data
in cache memory for each MN increases the system complexity.
2.4.6 User profile replication
This strategy replicates the user patterns in some suitable local
databases. When an incoming call arrives, the MN is searched in the
local database where the call is originated. If the called MN is found, the
VLR where it is roaming is obtained, and the call can be delivered. If the
called MN is not found, the GSM-MAP or IS-41 classical procedure is
followed. When a MN moves from a VLR service area to another, the
MN sends a Location Update message to maintain the location of the
MN. In paging process, the replication of the MN’s locations was
updated (obviously, the paging cost increases in comparison with the
GSM-MAP or IS-41 standards). This is also the difference between the
Per-user location caching scheme and the User profile replication
scheme. This has been dealt by Narayanan and Jennifer (1995). The
cache memory does not have to be updated when there is a probability of
a cache miss. This scheme can be very interesting depending on the ratio
between the incoming call and the mobility rates (per MN).
2.4.7 Connection Admission Control (CAC) Scheme
Fei et al (2005) observed that the basic idea of the CAC
scheme is to verify the feasibility of accepting new and handoff
connections under the conditions of guaranteeing the QoS of existing
connections and maximizing the utilization. The problem of LM is
usually divided into two parts: paging and location updating.
68
Both paging and update consume scarce resources like wireless
network bandwidth and mobile equipment power. These two important
areas have traditionally been addressed separately in the literature using
different sets of mobility information. Users may experience
performance degradations due to mobile handoffs. This problem
becomes more challenging in mobile networks. This problem has been
addressed using CAC.
In this scheme, each MN counts the number of boundary
crossings between cells incurred by its movements. A location update is
performed when this number exceeds a predefined movement threshold.
For simplicity, 1 is used as the threshold in the simulations. New
connection blocking probability and handoff dropping probability as
functions of offered load has been observed. Location update scheme is
movement-based which generates the movement path history.
2.4.7.1 Dynamic location management scheme
All available mobility information such as bandwidth units
(offered load), cell diameter, call-to-mobility ratio, in-session mobility
information, and cell residence time are used in this scheme. A common
mobility prediction scheme which can be used in both paging in MNs
and making admission decision. If the position of a user can always be
predicted in advance, then no explicit update is necessary.
2.4.8 Spatial-Quantization Based Update Algorithm
Akyildiz and Wang (2002) reported Global wireless networks
enabled mobile users to communicate regardless of their locations. One
69
of the most important issues was LM in a highly dynamic environment
because mobile users might roam between different wireless systems,
network operators, and geographical regions. A location-tracking
mechanism was introduced that consist of intersystem location updates
and intersystem paging. Intersystem update was implemented by using
the concept of boundary location area, which was determined by a
dynamic location update policy in which the velocity and the QoS were
taken into account on a per-user basis. Also, intersystem paging was
based on the concept of a boundary location register, which was used to
maintain the records of mobile users crossing the boundary of systems.
This mechanism not only reduced location-tracking costs, but also
significantly decreased call-loss rates and average-paging delays. The
performance evaluation of the proposed schemes was provided to
demonstrate their effectiveness in multitier personal communication
systems.
2.4.9 Effective utilization of system resources
Qi and Venkatasubramanian (2006) reported that efficient
resource provisioning that allowed for cost-effective enforcement of
application QoS relied on accurate system state information. However,
maintaining accurate information about available system resources was
complex and expensive, especially in mobile environments where system
conditions were highly dynamic. Resource provisioning mechanisms for
such dynamic environments able to tolerate vagueness in system state
while ensuring adequate QoS to the end-user.
In this work, authors addressed the information collection
problem for QoS-based services in mobile environments. Particularly,
authors proposed a family of information collection policies that vary in
70
the granularity at which system state information was represented and
maintained. They empirically evaluated the impact of these policies on
the performance of various resource provisioning strategies. They
generally observed that resource provisioning benefits significantly from
the customized information collection mechanisms that take advantage
of user mobility information. Furthermore, their performance results
indicated that effective utilization of coarse-grained user mobility
information (allowing multiple queries to run concurrently and system
will support higher throughput) renders better system performance than
using fine-grained user mobility information. (System attempts to
parallelize single tasks (example queries)). Using results from their
empirical studies, authors derived a set of rules that supported seamless
integration of information collection and resource provisioning
mechanisms for mobile environments. These results had been included
into an integrated middleware framework Automatic Service
Composition (AutoSeC) to provide support for dynamic service
brokering that ensured effective utilization of system resources over
wireless networks.
2.4.10 Trade-off between Location Update and Paging
Abhishek et al (2007) proposed two information-theoretic
techniques for efficiently trading off the location update and paging costs
associated with MM in wireless cellular networks. Previous location
tracking approaches always attempt to accurately convey a mobile’s
movement sequence. In these techniques, rate distortion theory is used to
arbitrarily reduce the update cost at the expense of an increase in the
corresponding paging overhead.
71
The goal of this paper is to design a common adaptive LM
technique that enables each individual MN to tune its mobility related
signaling load under a unified cellular access infrastructure. Such
individualized adaptation of the signaling load is becoming important
due to heterogeneity at various levels such as devices (automobiles,
personal medical monitors, laptops, cell phones, and so on), converged
applications like voice-over-IP, instant messaging, ondemand TV, and so
forth, and access technologies like 3Gcellular, 802.11, and WiMax.
Authors used the term “trade-off” since the paging and update
costs are mutually conflicting. In general, the precise degree of trade-off
depends on the application requirements and characteristics of individual
MNs. In this paper, authors imagined a user-specific LM framework
where the update and paging policies adapt to the movement and calling
patterns of individual devices. As telecommunications technologies
evolve, per-user schemes are expected to become more attractive,
especially if they can significantly reduce the signaling load across
billions of heterogeneous mobile nodes. In the nearer term, intelligent
per-user location tracking may be deployed in small and medium-sized
cellular networks such as smart homes/offices and campus Local Area
Networks (LANs). In this paper, authors had presented an effective
framework that trades off between the location update and paging costs
for wireless environments such as wide-area cellular networks and
extended campus networks. Their approach created per user movements,
in effect customizing the location update and paging sequence for each
user to its observed mobility pattern. When reduce the location updates,
the paging of the cellular network is increased. When paging cost
increases, the paging overhead and storage overhead also increase.
72
Simulation results demonstrated that the proposed algorithms
could reduce the update cost to very low values, with only modest
increases in the paging overhead. However, per-user profile may present
a scalability concern in tier-1 (national) cellular provider networks,
where the individual profile of millions of devices must be stored
separately. To make the schemes practical in large cellular environments,
future research could be investigated.
2.4.11 Quantitative analysis of the location management effect on
QoS
Yan et al (2007) stated that the fundamental component in
wireless network is, LM which consists of two operations: location
update and paging. These two supplementary operations enabled the
mobile user everywhere mobility. However, in case of failed location
update, a significant consequence was the obsolete location identity in
the network databases and thereafter the incapability in establishing the
valid route for the potential call connection, which would seriously
degrade the network QoS. In this work, authors performed a quantitative
analysis of the LM effect on QoS in the wireless networks. The metric
CBP adopted by Andrea et al (2011) and Zhaoji et al (2008) reports that
the average number of blocked calls was introduced to reflect the QoS.
For the sake of general applicability, the performance metrics were
formulated with the relaxed tele-traffic parameters. Namely, the call
inter-arrival time, cell residence time, location area residence time and
location update inter-arrival time followed a general probability density
function. The formulae were additionally specified in the static and
several dynamic location management mechanisms. Numerical examples
73
were presented to show the interaction between the performance metrics
and location management schemes. Yan et al (2007) analyzed theoretical
formulation and general modeling for the LM and QoS interaction.
The static LM is employed in the GSM/IS-41network. The
whole network is partitioned into several consecutive Location Areas
(LAs). Each LA is comprised of a collection of cells. The innermost cell
“0” is called the center cell or the cell in ring-0. Cells near the center cell
form the first ring around cell ”0”, cells near thering-1 form the second
ring around cell ”0” and so forth. The outermost cells are in the ring K.
In the static LM, whenever an MN enters a new LA, it will
perform the Location Update (LU) to register its new location to the
mobility databases. To reduce the associated signaling cost in updating
and paging, a variety of dynamic LM had been proposed. These include
timer-based scheme, movement-based scheme, distance based scheme
and pointer forwarding scheme.
However, LU requests may be failed and this phenomenon is
not a rare event due to the inherent characteristics of wireless systems,
e.g., physical link unpredictability, limited bandwidth and power
resources, and seamless mobility. In particular, the potential reasons
could be:
(1) The very limited bandwidth in Base Station (BS). In the
hot-spot locations, e.g. supermarket, bus interchange or public
transportation systems, a large number of travelers move into the same
area concurrently and trigger the LU procedure simultaneously.
However, due to the very limited bandwidth in a BS, only small part of
74
LU requests has the opportunity to capture a radio channel and succeeds
in refreshing the new information. The rest MN are unable to obtain a
radio channel and consequently the LU requests are rejected;
(2) The mobility database HLR or VLR failure. HLR stores the
subscribers’ permanent profile while VLR manages one LA and stores
the temporary records. In case, when the record in the HLR or VLR is
corrupted or outdated, LU rejection will occur,
(3) The limited space of VLR. When a group of passengers
enter an LA simultaneously, the responsible VLR will create new
records for these MNs. In case, the buffer is full, the VLR is not able to
create new record and the LU is failed,
(4) The unreliable wireless channel. One inherent
characteristic of wireless systems is the unreliable wireless link between
the Mobile Terminal (MT) and BS. The unreliability property or
potential link breakdown is involved in any functionality, including the
LU process.
The failed LU procedure has a direct impact on the call QoS.
Figure 2.11 Illustration of blocked call in case of failed LU
75
Figure 2.11 illustrates an example of the consequence of LU
failure. When MN1moves from the old LA to the new LA, MN1 will
send the LU message to VLRn. VLRn will then forward the LU message
to HLR to update the record by indicating the new LA identity.
However, due to some reasons mentioned above, the LU is failed. The
record in HLR is out-of-date, showing that MN1 is still residing in VLRo.
Suppose that MN2 wishes to make a call to MN1. Since the profile in
HLR shows that MN1 stays in VLRo, HLR will send a command to
VLRo to page MN1. However, at this time, MN1 is actually residing in
VLRn. Therefore, the system will not receive response from MN1 and no
valid routing can be established. The call request to MN1 is hence
blocked.
In this paper, authors theoretically and qualitatively analyzed
the interaction between the LU and QoS. Authors pointed out and
analytically investigated the LU failure and its consequence on the call
activity. The specific results in the basic location management and
dynamic schemes were developed. In addition, different from the
literatures those authors derived the distribution of the LA residence time
based on the network architecture and mobility model.
2.4.11.1 System Model for Call blocking probability
A. CBP due to LU failure
Figure 2.12 Time diagram to show the call blocking probability
76
Figure 2.12 illustrates the timing diagram of an MN’s behavior
taking in account the LU sequence and call activity. The symbols Li (i =
1, 2,..) denotes the instant that the MN performs the ith LU. Upon each
LU performing, the LU may be failed with probability Pf. The figure
illustrates that after (i 1) consecutive LU failure, the ith LU is
successful. tLU denotes the LU inter-arrival time with mean 1/ LU,
probability density function (pdf) (t),Cumulative Distribution
Function (CDF) FtLU(t) and Laplace Transform (LT) of
pdf (s).Ci(i=1,2,…) represents the moment that the ith call connection
request arrives to the MN. Let tc denote the call inter-arrival time with
mean 1/ c, pdf (t) and LT of pdf (s).Co represents the epoch of the
call arrival rate before the first LU. Let be the residual call inter-arrival
time, which terms the time duration from an intermediate moment to the
next call arrival. Then, the duration between the first LU and the
first call arrival follows the residual call inter-arrival time distribution.
Referring to the Residual Life Theorem proposed by Beakcheol and
Sichitiu (2012), the pdf of c is given by
( ) [1 )] with its LT ( ) = [1- (s)]/s.
After the first LU failure, if a call connection arrives between
T, this call is blocked. In other words, if is less than T, then the call is
blocked. Define the probability that a call request is blocked given that
an LU is failed. The probability is expressed as,
= Pr ( < T) = (t) (t) dt
77
B. Average Number of blocked call due to LU failure
The probability is able to exhibit the effect of the failed LU
upon one specific call connection. In case that the call activity is
frequent, the probability is incapable to reflect the frequency caused by
the LU failure. Authors denoted Mb as the number of blocked call
between T. For presentation, it is denoted tc,k equal to i.e. the
duration between the(k 1)th and kth call arrival.
Figure 2.13 Time diagram to show the number of blocked calls
Figure2.13 shows that there are Mb = k blocked calls after a
LU failure. Thereafter, the average number of blocked call between T is
formulated as
E ( )= (t)u(t)dt
=
= ))
78
C. Location Management schemes for Call blocking probability
For different LM schemes, the LU inter-arrival time tLU is
different. In this section, four schemes were discussed Basic Location
Management (BLM), Pointer Forwarding Scheme (PFS), Movement-
based Scheme (MBS) and Timer-based Scheme (TBS).
a. Basic Location Management (BLM):
Let denote the cell residence LA time with mean 1/ , pdf
(t) and LT of pdf (t). We term the LA residence time as the time
duration an MT stays in an LA. Let tLA be the residence time with pdf
(t) and LT of pdf (s). In the BLM employed by the GSM/IS-41
systems, whenever an MT enters a new LA, an LU is triggered. Hence,
the LU inter-arrival time tLU follows the LA residence time distribution
in this case.
b. Pointer Forwarding Scheme (PFS):
In the PFS, when an MT enters a new LA, the new VLR
exchanges messages with old VLR to setup a pointer to the new VLR.
The length of the chain of the forwarding pointers is limited to be L to
ensure the locating the MT within an acceptable delay. That is to say,
whenever the MT crosses L.LA boundaries, the MT will register itself
through the LU. Hence, the LU inter-arrival time is the summation of (L
+ 1) LA residence time. The LU arrival rate is .
c. Movement-Based Scheme:
Denote the LU threshold in MBS as M. Whenever an MT
experiences M cell boundary crossing, it will perform the LU process.
Hence, the LU inter-arrival time tLU is the summation of M cell residence
time. The LT of tLU is given by
79
(s) = )]
The LU arrival rate is expressed as
=
d. Timer-Based Scheme:
In TBS, every a fixed timer length Ttimer, the MT delivers LU to
the network. Hence, tLU is not a random variable, but a constant Ttimer.
The interval T is a discrete random variable. Its probability distribution
is,
Pr (T=k ) = (1- ); = 1,2 …
The CBP is given by
= Pr ( <T)
= Pr ( ) (1- )
The important QoS metrics CBP and the average number of
blocked calls are discussed under the generalized assumptions for the
tele-traffic parameters for the sake of broad applicability.
2.4.12 Buffer scheme for solving location management congestion
Yan and Fujise (2007) stated that the LM congestion problem
arises when the group of mobile nodes moves towards same location
area. In the hot-spot locations, e.g., supermarket, bus interchange, or
public transportation systems, a large number of travelers move into the
same area concurrently and trigger the LU procedure simultaneously to
register their new locations. However, due to the limited bandwidth in a
80
BS, only a small part of the LU requests had the opportunity to capture a
radio channel and succeeds in refreshing the new information to the
corresponding database in the wireless networks. The remaining MNs
were unable to obtain a radio channel, and therefore, the LU requests
were rejected.
A significant consequence of a failed LU is the out-of-date
location identity in the network databases and, thereafter, the incapability
in establishing the valid route for the potential call connection request,
which would seriously degrade the network QoS. Authors developed a
queuing model to characterize the congestion issue and proposed a
buffer scheme to eliminate this problem. The consequence of the LU
failure on the MN's call activity was also investigated. The comparison
demonstrated the high efficiency of the proposed approach in decreasing
the LU failure as well as the resulting CBP. The analytical model had
been validated by the discrete event simulation.
2.4.13 User Profile History based Location Management
Singh and Karnan (2010) observed that Cellular mobile
systems provides access to a wide range of services and allow MNs to
move randomly from one place to another within a well-defined
geographical area. Due to the growing number of mobile users, global
connectivity, and the small size of cells, one of the most critical issues
regarding these networks is location management. The challenging task
in a cellular system is to track the location of the mobile users effectively
so that the connection establishment cost and delay is low. In recent
years, several strategies had been proposed to improve the performance
of the LM procedure in mobile networks. In this work, authors proposed
81
an intelligent approach by taking the User Profile History (UPH) to
reduce the LU cost by combining Back-Propagation Algorithm and
Cascaded Correlation Neural Network. The implementation of this
strategy had been subject to extensive tests. The results confirmed the
efficiency of UPH by reducing the costs of both location updates and call
delivery procedures when compared to the various other strategies.
2.5 HANDOFF MANAGEMENT
Akyildiz et al (2007) observed that handoff (or handover)
management enabled the network to maintain a user’s connection as the
mobile terminal continues to move and change its Access Point (AP) to
the network. The three-stage process presented by Guangbin and
Jingyuan (2004) and Shantidev and Ian (2006) for handoff, first-stage
involves initiation, whether the user change network conditions or not
and identify the need for handoff. The second stage is new connection
resources for the handoff and performing any additional routing
operations. For Mobile-Controlled Handoff (MCHO), the MN finds the
new resources and the network approves. The final stage is data flow
control, where the delivery of the data from the old connection path to
the new connection path is maintained according to decided service
guarantees. The HM operations and research issues are shown in Figure
2.14(a) and 2.14(b).
82
Figure 2.14(a) Handoff Management operations Figure 2.14(b) Research Issues
HM includes two conditions, intra-cell handoff and inter-cell
handoff. Intra-cell handoff occurs when the user moves within a service
area (or cell) and experiences signal strength fall below a certain
threshold that results in the transfer of the user call to new radio channel
of appropriate strength at the same BS. Inter-cell handoff occurs when
the user moves into an adjacent cell and all of the terminal connections
must be transferred to a new BS. While performing handoff, the terminal
may connect to multiple BS simultaneously and use some form of
signaling diversity to combine the multiple signals. This is called soft
handoff. On the other hand, if the node stays connected to only one BS at
a time, clearing the connection with the former BS immediately before
or after establishing a connection with the target BS, then the process is
referred to as hard handoff.
HM research issues are i) efficient packet processing and
minimization of network signaling load ii) optimization of route for each
connection and efficient bandwidth reassignment in wireless
83
connections. Figure 2.14(b) lists the issues for the HM operations. In
future communications, networks will be considered the standardized
wireless network architecture in order to access regional, national, and
global services.
2.5.1 Distributed Bandwidth Management for QoS
Guohong (2001) observed that it was a challenge to support
QoS using limited frequency spectrum. In the literature, two orthogonal
approaches were used to address the bandwidth utilization issue and the
QoS provision issue; that is, channel allocation schemes had been
proposed to improve bandwidth efficiency. The HM schemes presented
by Shun-Fang and Jung-Shyr (2007) and Taleb and Letaief (2010) had
been proposed to guarantee a low connection dropping rate based on
bandwidth reservation. The authors integrated distributed channel
allocation dealt by Duy et al (2012) and adaptive HM dealt by Kim
(2011) to provide QoS guarantees and efficiently utilize the bandwidth.
Extensive simulations were provided and used to justify the analysis.
Compared to previous schemes, the proposed scheme could improve the
bandwidth utilization when providing QoS guarantees.
2.5.2 Cellular IP Protocol to reduce losing of data packets
The major objective of the work proposed by Jen-Shiun et al
(2003) was to propose multitier wireless communication architecture
based on Mobile IP and Cellular IP to support the service requirements
of mobile Internet and mobile multimedia communication. Based on this
architecture, there were some issues of research would be executed. How
the users’ requirement is satisfied for mobile Internet through wireless
communication.
84
In the proposed architecture, the overhead of system
management is decreased and the total effectiveness is improved.
Furthermore, the handoff and location management methods are
presented to improve QoS. By the way, resource switching management
was introduced to reduce data packet loss.
2.5.2.1 Multi-tier Cellular Architecture overview
Mobile Internet architecture is considered including an overlap
hierarchical framework. Each framework has its individual feature, i.e.
satellite, macro cell, microcell and pico-cell area used by Hoang et al
(2000) and Karapantazis and Pavlidou (2005). Such that, by applying
this framework, it supports different transfer rates between mobile nodes
and distinct geographical areas. Figure 2.15 demonstrates the framework
of Multi-tier cellular architecture.
Figure2.15 Multi-tier Cellular Architecture
Mobile IP was optimized for macro-mobility and relatively slow-
moving MNs.
85
Cellular IP represented a new mobile node protocol that is
optimized to provide access a mobile IP enabled Internet in
support of fast moving wireless nodes. It could offer fast handoff,
less delay, a few or even no packet loss between base stations.
2.5.2.2 Handoff Strategy
This multi-tier architecture defines a domain to be coverage of
macro-tier. Hence, handoff strategies can be distinguished into Inter-
domain handoff and Intra-domain handoff:
Inter-domain Handoff
When MN moves from one domain to the other, Inter-domain
handoff is happened and it has two following situations:
a. The upper layer BS of these two domains is same
As shown in Figure2.16, when MN moves to a new domain, it
needs to ask the BS of macro-tier for handoff. If macro-tier has no free
channels for handoff, MN turns to ask micro-tier for handoff. MN will
send a location message to R3 through micro-tier or macro-tier BS to
update its location information after successful handoff.
Figure 2.16 the upper BS of two domains are same
86
b. The upper layer BS of these two domains is different
While new domain permits handoff, as shown in Figure 2.17,
MN sends an update location-message to new macro-tier, and macro-tier
would send this message to it supper layer. Due to the upper layer BS of
these two domains is different; the most upper layer BS needs to deliver
this message to home network of MN. Then, home network will reply
new location information to original domain. However, this record will
keep a while until MN has completed handoff.
Figure 2.17 the upper BS of two domains are different
Intra-domain Handoff:
Due to the proposed multi-tier network architecture is
consisting by macro-tier and micro-tier, Intra-domain Handoff can be
separated into three conditions as follows:
87
a. Macro-cell to micro-cell:
When MN moves to the area that macro-cell and micro-cell are
overlapping or while MN needs more bandwidth when it was served by a
macro-cell, system will switch MN to micro-cell.
If MN demands more bandwidth, it must wait system to accept
its request and then start switch to micro-cell. Furthermore, it must send
an “Update Location Message” to new BS and a “Delete Location
Message” to old BS in the same time.
b. Micro-cell to macro-cell:
As shown in Figure2.18, when MN Y moves to the areas that
do not cover by micro-cell, it needs switching to the BS of macro-cell.
Therefore, MN will send a handoff request message first. If system
accepts its request, it will send an “Update Location Message” to the BS
of macro-cell to store location information of MN in macro table. It will
forward the message to update the macro table of its parent macro-cell
BS.
c. Micro-cell to micro-cell:
When MN Z (see Figure2.18) moves from one micro-cell F to
the other E, as long as arrival at area that needs to demand a handoff
request, it musts send a request message to new BS. After new BS
accepts its request, it will send an “Update Location Message” to D and
modify the record of micro table. If there are no enough bandwidths of
micro-cell, it will turn to macro-cell for a handoff request. In this
situation, the handoff procedure is same as case b.
88
Figure 2.18 three situations of Intra-domain Handoff
2.5.3 Self-Adaptive Handoff Management
Bellavistaet al (2009) reported that self-adaptive management
and quality adaptation of multimedia services were open challenges in
the heterogeneous wireless Internet, where different wireless access
points potentially enabled anywhere anytime Internet connectivity. One
of the most challenging issues was to guarantee streaming continuity
with maximum quality, despite possible handoffs at multimedia
provisioning time. To enable HM to self-adapt to specific application
requirements with minimum resource consumption, this work offered
three main contributions.
First, it proposed a simple way to specify handoff-related
service-level objectives that were focused on quality metrics and
tolerable delay. Second, it presented how to automatically derive from
these objectives a set of parameters to guide system-level configuration
89
about handoff strategies and dynamic buffer tuning. Third, it described
the design and implementation of a novel handoff management
infrastructure for maximizing streaming quality while minimizing
resource consumption. Their infrastructure exploited i) experimentally
evaluated tuning diagrams for resource management and ii) handoff
prediction/awareness. The reported results show the effectiveness of
their approach, which permitted to achieve the desired quality-delay
tradeoff in common Internet deployment environments, even in presence
of vertical handoffs.
2.6 EVALUATION OF CALL DROPPING PROBABILITIES
Faultless user mobility in a wireless network is achieved
through the process of handoff in which a user’s signal is transferred
from one base station to another as the mobile user moves across the cell
boundary or when the link quality is unacceptable, enabling the mobile
node to roam around the service area. The number of channels allocated
to a base station determined the number of users that could be served by
that base station. When all the channels of a base station were assigned
to existing calls, newly arrived calls would be blocked, and handoff calls
from other base stations would be forced to terminate. Since terminating
ongoing calls was less tolerable than blocking a new call, a properly
designed handoff mechanism reduced the probability of blocking new
calls, when maintaining the QoS of ongoing calls. Frequent handoffs
reduced the QoS, increased the signaling overhead and degrade data
throughput in the network as demonstrated by Beakcheol and Sichitiu
(2012). A handoff decision algorithm might be made based on the
received signal strength, bit or frame error rate measurement, distance
90
between mobile user and the base station, traffic load, mobile velocity,
etc. A predictable handoffs decision scheme compared the received
signal strength from one serving base station with that from a target base
station, using a handoff threshold dealt by Hwa-Chun and Show-Shiow
(2000)which must be carefully determined to minimize hysteresis.
2.6.1 Homogeneous wireless Network
Aalo and Efthymoglou (2010) observed that in wireless
networks, tele-traffic parameters that affected the achieved network
performance include the Call Holding Time (CHT) and the Cell Dwell
Time (CDT). The first specified the time duration that a user stays
connected to the network and depends on the type of service required,
whereas the CDT is the time that the mobile user resides in a cell, and it
is related to the user mobility and the cell type. The CDT is a mobility-
related parameter that affects the network performance, since every time
a handoff occurs there is a distinct probability p, that the call may be
suddenly terminated due to unavailable radio resources in the new cell.
In a cellular network, one common theme is the cells are assumed to be
identical and, therefore, the probability of handoff failure in each cell
was constant. However, in practice, the cellular footprint and the traffic
in each cell are usually irregular. The authors presented results on call
performance metrics for homogeneous networks.
91
2.6.1.1 Performance on Homogeneous Wireless Network
Figure 2.19 Time diagram for a complete call
Let the random variable ‘ ’ denoted the CHT, i.e. the duration
of time from the time a call is initiated and successfully connected to a
cellular base station to the time the call is terminated by the mobile user.
Also let Tk, k=1, 2, 3,… denote the duration of time that a mobile
terminal resides in the k-th cell. The time diagram of a complete call is
depicted in Figure 2.19.
From Figure 2.19, it is observed that in a wireless network a
call ends randomly in the K-th cell, where K is a random variable. For a
homogeneous cellular network, K is a geometric random variable with
parameter p since the call successfully completes(K-1) handoffs, each
with probability (1 p) and fails on the K-th handoffs attempt with
probability p , i.e.,
Pr ( | ) = (1 ) , k = 1, 2, 3,…
92
Let p0 denote the probability that a new call is blocked when trying to
connect to the wireless network. Furthermore, let pD be the CDP, i.e. the
probability that a new call which is successfully connected to the cellular
network is subsequently dropped when a handoffs attempt fails.
Therefore the CDP may be defined as
pD = Pr (the call is not blocked and it is forced to terminate due
to unsuccessful handoff)= ( ) Pr( )where Sk = R1 +T2 +T3
+…. +Tk is a random sum of independent non negative random variables
[9]. Authors assumed that the Tk are i.i.d. random variables with
common probability density function (pdf) fT(t).
Authors validated the analytical expressions and illustrate the
effect of the call-to-mobility factor , on the call dropping and call
completion probabilities. In all numerical results, authors plotted the call
Dropping or call completion probability versus the call-to-mobility
factor, for homogeneous networks. They assumed exponentially
distributed CHT with = 1/5. For the CDTs, it was assumed that (1, 2,
3,..., k,...) are independent and identically distributed (i.i.d.), exponential
variables with parameter . The call-to-mobility factor is defined as the
ratio of the mean of the call holding time to the mean of the cell dwell
time. i.e., = ( )( ) = .
From this equation the value of can be obtained for each
value of . using these random variables for the CHT and the CDTs a
simulation model implementing the communications scenario depicted in
Figure 2.19 was created in Matlab. Figure 2.20 plots both the analytical
and simulation results of the CDP assuming constant probability of
handoff failure as well as random probability of handoff failure. For
93
comparison purposes, authors assumed the constant value of p to be half
of the upper limit of the uniform distribution. From this figure, authors
observed the performance improvement when having random probability
of handoff failure in each cell compared to a constant value.
Furthermore, the performance difference between the random and
constant value cases increases for higher values of and higher values of
call-to-mobility ratios. Similar observations can be made from
Figure2.21 in terms of the call completion probability, which decreases
with the increase of and , as expected.
Figure 2.20 pD vs for constant and random values of p.
94
Figure 2.21 pc vs for constant and random values of p
In this work authors derived analytical expressions for the call
dropping and call completion probabilities in a heterogeneous wireless
network where the handoff failure in each cell is not constant but it
follows a uniform distribution. Simulation results were shown to agree
with the analytical results, thus verifying the correctness of the analytical
expressions.
Merits: Reduced call dropping probability
Demerits: 1. Increased the signaling overhead and
2. Degrade data throughput in the network
2.7 NEXT GENERATION VALUEADDED MOBILE SERVICES
Zaghloul et al (2009) observed that recently, all-IP wireless
systems had been standardized under the IP Multimedia Subsystem
(IMS) framework to support next generation value added mobile
95
services. In all-IP networks, multiple base stations connect to IMS policy
servers through IP-based access gateways. As users moved from one
access gateway area into another, the corresponding access gateways
initiated signaling flows for QoS authorization toward the IMS policy
server. Depending on the service, the policy function might contact one
or more application servers resulting in variable and expanded signaling
delays upon each handoff. Such delays could be of the order of seconds,
depending on the specific system design and whether the user was
roaming, which might be critical to the quality of real time services. In
this work, authors proposed a novel proactive signaling method in the
application layer that conveyed authorization delay constraints from the
IMS to the radio layer and thus mitigates the effects of variable signaling
delay as dealt by Bokrae et al (2010). Their method was practically
relevant as it uses the already established mechanisms of authentication
and authorization signaling via standardized interfaces and protocols.
Using the OPNET simulator, they demonstrated that their scheme was
also scalable, as its corresponding signaling overhead was upper
bounded to approximately double the handoff rate.
2.8 OPTIMAL EFFECTIVE BANDWIDTHPOLICY FOR CALL
ADMISSION CONTROL (CAC)
Bo et al (2008) reported that in Wireless Mesh Network (WMN),
it was important to provide an efficient handoff scheme, due to the
frequent user mobility. To address this issue, authors proposed a Mobile
Agent (MA) based handoff approach, where each mesh client had a MA
residing on its registered mesh router. To guarantee QoS and achieve
differentiated priorities during the handoff; Falowo (2010), Leong et al
96
(2006), Ling et al (2009) and Mohamed and Deniz (2004) developed a
proportional threshold structured optimal effective bandwidth policy for
CAC on the mesh router. Simulation study showed that their proposed
CAC scheme could obtain satisfying tradeoff between differentiated
priorities and statistical effective bandwidth in Wireless Mesh Network
handoff environment.
2.9CALL BLOCKING PROBABILITYWITH HIGH SPEED MOVING TERMINAL
Ioannou et al (2002) presented a two-layer mobile network
architecture which optimized the handoff blocking probability
performance of High-Speed Moving Terminals (HSMT) in a congested
urban area. The lower layer of the proposed architecture was based on a
micro-cellular solution, for absorbing the traffic loads of both the Low
Speed Moving Terminals (LSMT) and the new calls of HSMT. The
higher layer was based on a macro-cell umbrella solution, for absorbing
the traffic load of the existed handoff calls of the HSMT. The results
showed that using the optimum number of channels in each layer, the
handoff CBP of the HSMT was optimized having the minimum effect on
the call blocking probability of the micro-cellular layer.
2.10 DYNAMIC LOAD BALANCING SCHEME
Yanmaz and Tonguz (2005) proposed to use a dynamic load
balancing scheme that utilizes fixed relay stations placed in a cellular
geographical coverage area to improve the handover performance in
cellular networks. To this end, authors developed closed-form
performance expressions and derive the handoff dropping probability in
terms of the main system parameters, such as the new call and handoff
97
arrival rates, the number of fixed and load-balancing channels per cell,
the load balancing probability for new and handoff calls, etc. Their
results showed that by employing a dynamic load balancing scheme, not
only the new CBP, but also the handover dropping probability could be
improved considerably, without reserving any channels or assigning any
priority to handoff calls. In fact, employing dynamic load balancing
without any guard channels outperforms a conventional system that
employs the well-known Guard Channel Method (GCM).
2.11 HYBRID CHANNEL ALLOCATION (HCA) SCHEME
Joshi and Mundada (2010) describes that in wireless mobile
communication systems, the radio spectrum is limited resource.
However, efficient use of such limited spectrum becomes more
important when the two, three or more cells in the network become hot-
spot. The use of available channels had been shown to improve the
system capacity. The role of channel assignment scheme was to allocate
channels to cells in such way as to minimize call-blocking probability or
call dropping probability and also maximize the quality of service. In
this work attempts were made to reduce call-blocking probability by
designing Hybrid Channel Allocation (HCA) which was the combination
of Fixed Channel Allocation (FCA) and Dynamic Channel Allocation
(DCA). A cell becomes hot-spot when the bandwidth available in that
cell was not enough to sustain the users demand and call would be
blocked or dropped.
A simulation result showed that HCA scheme significantly
reduced call-blocking probability in hotspot scenario and compared with
cold-spot cell. This hot-spot notification would request more than one
98
channel assigned to the requesting cell, proportional to the current hot-
spot level of the cell. Furthermore, all channels would be placed in a
central pool and on demand would be assigned to the base station. That
would be helpful to reduce call-blocking probability when cell becomes
hot-spot. When a call using such a borrowed channel terminates, the cell
might retain the channel depending upon its current hot-spot level
therefore HCA had comparatively much smaller number of reallocations
than other schemes. It also showed that it behaved similar to the FCA at
high traffic and to the DCA at low traffic loads as it was designed to
meet the advantages of both.
2.12 HANDOFFANDCDPIN WIRELESS CELLULAR
NETWORKS
Iraqi and Baoutaba (2005) describes that in cellular networks,
the CDP was a very important connection-level QoS parameter. It
represented the probability that a call is dropped due to a handoff failure.
The goal of almost all admission control schemes was to limit the CDP
to some target value while maintaining higher bandwidth utilization or
lower blocking rates for new calls in the system. Another related
parameter was the Handoff Dropping Probability (HDP). It represented
the probability of a handoff failure due to insufficient available resources
in the target cell. Most local admission control schemes try to limit the
HDP to some target maximum and assume that this limits the CDP too.
In this work, authors showed that even if the HDP is controlled to be
below a maximum value in every cell in the network, the CDP
experienced by the users is not controlled, independently from the
admission control scheme used to control the HDP.
99
2.13EVALUATION OF CALL PERFORMANCE IN CELLULAR
NETWORKS
The unstable demand for wireless services in recent years had
led to the development and the deployment of a new generation of
wireless systems that offer a variety of high-speed wireless applications,
including voice and data services, multimedia services, navigation
services, text-video messaging, and Internet browsing.
To satisfy the QoS requirements, these systems encounter
problems that are related to the user’s mobility in the cell, i.e., outage
due to fading over the wireless link and handoff failure due to the
shortage of spectral resources as the user moves through the coverage
area of the cellular system. In a wireless network, the instantaneous
received signal suffers from short term fading, which is usually modeled
by the probability density functions (pdfs), and from the slow variation
of its local mean power (known as long-term fading) that is usually
modeled by the lognormal distribution. A parameter that is commonly
used to quantify the effect of channel fading on the performance of a
wireless network is the minimum time (duration) that the received signal
level stays below a preset threshold, causing the call to be dropped by
the base station. The minimum duration for an outage event is derived
using a level crossing analysis, and it usually spans widely different
timescales for the short and long term fading channels.
In a cellular system, the coverage area is divided into cells, and
a portion of the total available channels is assigned in each cell.
Neighboring cells are assigned different groups of channels to minimize
100
the interference between cells. Two time dependent random parameters
that are usually encountered in the study of cellular network performance
are the cell dwell time and the call-holding time. The cell dwell time of a
mobile user is a random variable that characterizes the time that the
mobile user spends in a cell, and it depends on many factors, including
the cell size, the terrain, the speed, and the direction of the user.
Although it has usually been assumed to follow an exponential
distribution, more generalized distributions that provide better fit to
measured data have recently been employed. Furthermore, the call-
holding time mainly depends on the call type, i.e., voice, data, Web
browsing, etc. Although, for voice calls, this random variable had been
assumed in the past to follow an exponential distribution, more
generalized distributions had been proposed for the new types of calls in
the new generation of wireless networks.
The call performance analysis of homogeneous wireless
networks under the generalized cell dwell time and call-holding time had
been studied. This approach was based on the residue theorem and
requires that the Laplace transform of the distribution of the call-holding
time be a rational function. However, in some cases of practical interest,
the Laplace transform of the call-holding time distribution may not be a
rational function as described by Pattaramalai et al (2009). For example,
when the call-holding time has a gamma distribution with a non-integer
shape parameter, its Laplace transform does not have a rational form,
and the residue theorem approach is not applicable.
A new approach based on the concept of random sums had
been proposed by Pattaramalai et al (2009), Call-completion probability,
101
CDP, and handoff rate are important performance measures of wireless
networks. Authors studied the joint effect of channel fading and handoff
failure on these performance measures. For the case of Rayleigh and
lognormal fading channels and for generalized distributions of the cell
dwell time and the call-holding time, authors derived simple closed-form
expressions that closely approximated these performance metrics. The
results were given in terms of the Moment Generating Function (MGF)
of the distribution for the call-holding time and might be useful in the
cross-layer design and the optimization of wireless networks.
2.14 CDPWITH UNIFORMLYDISTRIBUTED FAILURE RATE
Aalo and Efthymoglou (2010) derived closed-form expressions
for the CDP and call completion probability for a mobile network with
general distributions for the cell dwell time. In particular, authors
considered a non-homogeneous mobile network in which the handoff
request failure in each cell followed a uniform distribution. Authors
assumed throughout this paper that the channel holding time was
exponentially distributed but the cell dwell time had an arbitrary
distribution. For special cases of the cell dwell time distribution, closed-
form expressions for the call performance metrics were obtained.
2.15 CELL SPLITTING IN MOBILE NETWORKS
Theodore (2007) explained that cell splitting is the process of
subdividing a congested cell into smaller cells, each with its own base
station and a corresponding reduction in antenna height and transmitter
power. Cell splitting increased the capacity of a cellular system since it
102
increased the number of times that channels were reused. By defining
new cells which have a smaller radius than the original cells and by
installing these smaller cells (called micro-cells) between the existing
cells, capacity increased due to the additional number of channels per
unit area.
Figure 2.22 Illustration of Cell splitting
An example of cell splitting is shown in Figure 2.22. The BSs
are placed at corners of the cells, and the area served by base station ‘A’
is assumed to be saturated with traffic (i.e., the blocking of base station
‘A’ exceeds acceptable rates). New base stations are therefore needed in
the region to increase the number of channels in the area and to reduce
the area served by the single base station. Note in the figure that the
original base station ‘A’ has been surrounded by six new micro-cells. In
the example shown in Figure 2.22, the smaller cells were added in such a
way as to preserve the frequency reuse plan of the system observed by
Koyluoglu and El (2009) and Novlan et al (2011). For example, the
micro-cell base station labelled ‘G’ was placed half way between two
larger stations utilizing the same channel set ‘G’. This is also the case for
103
the other micro-cells in the figure. As it can be seen from Figure 2.22,
cell splitting merely scales the geometry of the cluster. In this case, the
radius of each new micro-cell is half that of the original cell.
For the new cells to be smaller in size, the transmit power of
these cells must be reduced. The transmit power of the new cells with
radius half that of the original cells can be found by examining the
received power Pr at the new and old cell boundaries and setting them
equal to each other. This is necessary to ensure that the frequency reuse
plan for the new micro-cells behaves exactly as for the original cells.
2.16 SUMMARIZATION OF LOCATION MANAGEMENT AND
HANDOFF MANAGEMENT TECHNIQUES
The features of LM and HM are summarized and are given in
table with pros and cons.
Table 2.2 Summarization of Location Management Techniques
S.No. Techniques Features Pros Cons
1. HLR/VLR Spatialquantization isused forregistration of amobile node.
The locationupdation cost is low.
i) The locationupdation delay ishigh.ii) The call droppingrate is high.
2. IP Micromobility
Local MobilityManagementServer (LMMS)storing theinformation ofroaming hosts
i) QoS wasguaranteed by staticLabel Switched Path(LSP)ii) Call dropping isless than otherprotocols
i) Systemmaintenance cost ishigh since severalpaging servers (PSs)included.
104
Table 2.2 Summarization of Location Management Techniques
(Continued)3. IS-
41(InterimStandard 41)
supportinteroperabilitybetween differingnetworks formobile telephony
Call delivery, SMSdelivery, validationand authenticationthrough an inter-system messagingprotocol are carriedout.
i) Location updationdelay is high.ii) Call droppingrate is high.
4. DynamicLocationAreamanagement
Calculating theoptimal LocationArea size giventhe respectivecosts for locationupdate.
ii) Call blocking rateis reduced.
i)Theimplementation ofthis scheme iscomplicated whencells are hexagonalshape, or irregularshape.
5. SpatialUpdation
i) Reliablelocation updationcostii) Most of themobile serviceproviders usingthis technique
i) The updation costis less.
i) The call droppingrate is high.ii)There is noreliable technique toreduce locationupdatesiii)Unnecessarylocation updatespossible
6. TemporalUpdation
The locationupdation iscontrolled bymobile serviceproviders.
i) It is possible toremove unnecessarylocation updations.ii)Packet deliveryratio is highiii) Repetition oflocation updationiii) Routingoverhead is low
i) Network signalingcost increases due tofrequent locationupdation.
7. Bufferscheme
It solves locationmanagementcongestion.
Call blocking rate isless
There is no efficientdynamic loadbalancing technique.
8. User ProfileHistory(UPH)scheme
Supports topredict thelocationupdation.
Location updationcost is low.
Sometimes wrongprediction of thelocation is possible.
105
Table 2.3 Summarization of Handoff management techniquesS.
No. Techniques Features Pros Cons
1. Mobile IPwith IP QoSmechanism
i) Supportmobility in IPnetworksii) Supportsoft handoff
i)Dynamic locationtrackingii) Call droppingrate is low.
i) Routing overhead ishigh.
2. MaISAM forfast mobility
Any type ofserviceanytime andanywhere
Call dropping rate isminimum.
High data ratesmultimedia service formobile user is notpossible.
3. NSIS( NextStep InSignaling)
SupportsMaISAM
Due to fast and asmooth handoff, calldropping rate is less.
No reliable handoff.
4. FHMIPv6 i) Supportmobility in IPnetworksii) Supportsoft handoff
i) Call dropping isminimumii) Handoff delay isminimum
Call blocking ratecannot be reduced.
5. ConnectionTree for IPmicromobility
i) Support softhandoffii)Fixed CoA
Reduce routingoverhead.
i) No scheme to reducecall blocking rate.ii) A mobile hold morethan one channel whenhand off takes place.iii) Shortages ofchannel service
6. Mobile IPFastAuthentication (MIFA)protocol
Simplenetworkarchitecturewithoutintroducingnew entities.
Call dropping rate isminimum.
Call blocking ratecannot be reduced.