A Survey on Cost Effective Survivable Network Design in Wireless Access Network

International Journal of Computer Science & Engineering Survey (IJCSES) Vol.5, No.1, February 2014
DOI : 10.5121/ijcses.2014.5102 11
A Survey on Cost Effective Survivable Network
Design in Wireless Access Network
K.Rajalakshmi and Krishna Gopal
Department of Computer Science, Jaypee Institute of Information Technology, Noida
ABSTRACT
In today’s technology, the essential property for wireless communication network is to exhibit as a
dependable network. The dependability network incorporates the property like availability, reliability and
survivability. Although these factors are well taken care by protocol for wired network, still there exists
huge lack of efficacy for wireless network. Further, the wireless access network is more complicated with
difficulties like frequencies allocation, quality of services, user requests. Adding to it, the wireless access
network is severely vulnerable to link and node failures. Therefore, the survivability in wireless access
network is very important factor to be considered will performing wireless network designing. This paper
focuses on discussion of survivability in wireless access network. Capability of a wireless access network to
perform its dedicated accessibility services even in case of infrastructure failure is known as survivability.
Given available capacity, connectivity and reliability the survivable problem in hierarchical network is to
minimize the overall connection cost for multiple requests. The various failure scenario of wireless access
network as existing in literature is been explored. The existing survivability models for access network like
shared link, multi homing, overlay network, sonnet ring, and multimodal devices are discussed in detail
here. Further comparison between various existing survivability solutions is also tabulated.
KEYWORDS
survivability, cost, capacity, optimization, wireless access network.
1. Introduction
Rapidly growing wireless network and increasing demands mobile communication attracts
research and development with emphasis on mobile data services. In current scenarios, the mobile
users utilize untethered access to wireless network resources. Wireless access network includes
personal communication services (PCS), cellular network and wireless local area networks
(WLAN). Wireless networks are more attractive in spite of low quality and reduced services than
wired networks. Research in wireless access network is to provide communication services at
“any time, any place, any form”. The vision of wireless access network is to provide services like
voice, data and multimedia to the users regardless of location, user mobility and type of access
terminal used. Users demand reliable services similar to wired telecommunication and data
networks which lead to intolerance to communication failure or critical data loss. Ability of a
network to perform its intended services upon infrastructure component failures is known as
survivability. Survivability is achieved by simple fault tolerance future. Failures occurs in
components like switches, base stations, databases, mobile devices, wired / wireless links among
mobile terminal, base station, switches. Wireless access network has tree like topology, is very

12
much vulnerable to failures. The nodes of the access network topology are BSs, BSCs, and
MSCs. The link or edges of topology are the physical transmission cables that establish
communication between nodes. Two types of failures occur in wireless access network, they are
node failure and link failure. Thus, survivability of the network in case of failure must be
addressed while designing access network. Even a single link failure could cost heavy damage to
the network services. Especially loss of critical communication cannot be tolerated and need an
effective restoration mechanism.
A generic wireless access network consists of cells base station, base station controllers and
mobile switch centers. Geographical service area of cellular network is partitioned into small
structures called cells. Each cell has a fixed base station which is access points for all mobile
terminals (MTs). Base station controller (BSC) manages group of Base stations (BSs). The
wireless links between the BS and MTs within cells are digital and employ time division multiple
access (TDMA) or spread spectrum code division multiple access (CDMA). BSs and BSCs are
connected to backbone network via Mobile switching centers (MSC). MSCs are connected to the
core network through access routers. Mobile switching center connects to the signaling network
further through transmission network. Associated with signaling network and MSCs are databases
that support users with service mobility like authentication and roaming. In general wireless
access network is associated with Home Location Register (HLR), Visitor Location Register
(VLR), Equipment Identity Registers (EIR) and Authentication centers (AUCs). The home
location register provides user profile information like types of services subscribed, billing
information and location information. The visitors’ location register provides information about
the mobile user’s visiting an associated MSC coverage area. Figure 1 depicts, various three level
of wireless access network namely BS level, BSC level and MSC level. At BS level the mobile
terminals communicate to their corresponding cell BSs. At BSC level, BSs connects to their
allocated BSCs. And at the MSC level BSCs connect to their corresponding MSCs. The
communication link between the BS, BSC and MSC are typically wire line or fixed microwave
links.
As given by Kubat et al. in [2] and also in [7], the cells originating and/or terminating at the
mobile units are connected to a cell site over a radio link (analog, digital). From the cell, the call
proceeds to the MSC for processing and switching. The switch transfers the call to its destination
– either another mobile unit, or to a PSTN for delivery to a wired telephone terminal. Between the
cell site and the MSC, the connection is over terrestrial, T1 digital transmission lines or over
private digital microwave.
Fig 1. A basic wireless access network

13
As stated by Houeto et al. [5], generally three types of networks are distinguished in the 3G
wireless networks. The core network is a WAN or a MAN, whereas the access networks or LAN
or WLAN which essentially consist of access routers (AR), hubs and Multiplexers. An access
network can have a star, tree, ring or mesh topology. In star topology; central master node
controls all the auxiliary nodes. The auxiliary nodes must communicate each other only through
the central master node. In case of tree topology; the principal node is the root node through
which all the communications takes place. The tree topology is a hierarchical structure with child
nodes. In case of ring topology, the nodes are connected in a point to point to form are further
interconnected to form a ring. Considering two nodes, they perform point to point communication
whereas a whole network, the nodes form ring topology. In case of fully meshed topology, each
node is connected to every other node. In partially meshed topology, each node is connected to
more than one node.
2. Need for survivability
The major objective to achieve end-end performance of networks and systems is dependability.
The dependable networks are valued based on three parameters namely reliability, survivability
and availability. Reliability is defined as the estimation of network failure probability. The
probability is measured using mean time between failures (MTBF). Availability is defined as the
estimation of network ready time for usage without any interference. The parameters like mean
time to repair, mean time between failures are used to estimate the availability. The objective of
availability is to achieve faster MTR and lesser MTBF which will improve the performance
tremendously. Survivability is defined as the ability of network to perform its assigned function
even during the event of failures of network components. The dependability in wired network like
local area network, wide area network have gained performance enhancement in recent years.
Further high speed networks like public switched network, internet, ATM, frame relay are been
researched very well. However, in case of wireless network, dependability is still requires many
issues to be addressed. One such essential parameter to be address in wireless network
performance is fault tolerance in the network components.
Varshney [3][6] discusses one of the major reasons for targeting survivability problem is to
develop an efficient dependable wired network. The dependable wired network incorporates all
three properties like Reliability, Availability and Survivability. Some dependable networks like
internet, public switched networks, ATM based networks although lacks proper enhancement
towards reliability, survivability and availability but has achieved these requirements to some
extent. However, the wireless networks lacks tremendous by the factors of survivability,
reliability and availability. The factors affecting these enhancements are diverse link types for
wireless medium like Bluetooth, Zigbee , RF etc, further adding to this are the factors like user
mobility , lack of regulations, high CAPAX and OPEX for wireless access networks. Also due to
technologies diversities like GSM, UMTS, CDMA, 3G LTE environments, the performance of
such wireless networks degrades.
Some of notable recent works are discussed further. Oriol et al [8] provided a roadmap for UMTS
optimization and illustrated a methodology to provide sub optimal coverage of cells. Gordejuela
et al., [9] described a way to deal with the tradeoff generated during the LTE access network
design, and presented a service oriented optimization framework that would offer a new
perspective considering technical and economical factor of cellular network. Quintero et al. [10]
targeted assignment of NodeBs to RNCs in large scale 3G cellular networks. Hybrid genetic

14
algorithm with migration was used to solve the NP hard problem. The authors propose multi
population approach where each population was partitioned into subpopulations. These works
further neccesiaties the research foucs on the end-end performance of wireless networks.
3. Failure Scenarios
Survivability strategies designed and analyzed for wired networks are not applicable for the
wireless network, major reasons being the user mobility and power management in the wireless
network. The end to end performance of wireless network is fully graded based on the user
mobility and power management. In order to facilitate survivable analysis and design of wireless
access networks, Tipper et al. [1] developed a survivability framework similar to the approaches
for wired backbone networks. The survivability framework for wireless access networks consists
of three layers with survivability strategies possible at each layer. The three layers are termed
access, transport, and intelligent. Note that the logical layers defined here are independent of the
physical implementation of the network. Each of the three layers is characterized by network
functions, network components, and communication links.
Fig 2. Failures in wireless networks
Generic example of failure scenarios in cellular / PCS network are (i) failure of a base station (ii)
loss of a mobile switching centre (iii) loss of link between a base station and a mobile switching
center. Metrics used to access survivability of network (i) Network performance using call
blocking probability (ii) traffic restoration efficiency based on percentage of failed demand been
restored. The failure of the wireless network components can affect the performance of networks
in several ways.

15
Figure 3. Impact of failures in inter communication between wireless cellular networks
This is illustrated using figure 3. Network X failure may occur due to any of the failures such as
MSC failure, BSC/BS failure, link failure and HLR/VLR failure. When such failure occur in
network X, the customer belonging to network Y is not able to access the roaming facilities of
network X. Similarly, considering network Z, utilizing group membership facility of network X,
but then as the network X fails due to various failure scenario, the group services are not available
for the network Z.
4. Survivable Strategies
Kubat et al. [2] and Dutta et al. [7] models the network design of cellular network as (i) Model #1
is a 1-period fixed-link capacity model (ii) Model #2 allows the capacities of the links to vary
(iii) Model #3 presents a multi-period demand problem. In this, ring type of network topology is
consider as restriction and analysis of optimal homing of the network nodes are been carried out.
Here each cell can home multi-homing. That is, the cells can be connected to different nodes. The
major restriction is that each of these nodes has to be on different disjoined trees. The above
restriction adds complexity to the network designing. By varying the capacity of links connecting
the nodes with each other, the performance has been evaluated. Considering the traffic flow, the
traffic comprises of cellular traffic and backbone traffic. The backbone traffic constitutes the
major traffic part the access network. Therefore, while performing network designing this
diversified load on the transmission link must be consider. Varshney et al. [3] [6] presents a
network design scheme using ‘selective’ multilevel fault tolerance for dependable infrastructure-
oriented wireless networks. The selective multi-level fault tolerance incorporates adaptable
building blocks (ABB). The diversified dependable networks with criteria like survivability,
reliability and availability are targeted by ABB using several levels of repairs in order to reduce
redundancy of network components. Te various levels of possible repairs are at switches, base
stations, BS controllers, and HLR and VLR databases. Tipper et al. [1] applies survivability
strategies at three different layers namely (1) at access layer (2) at transport layer (3) at intelligent
layer. At the radio level of access layer, the wireless link failure to the user is the primary failure.
To provide survivability during link failure, each base station supports two types of radio

16
channels namely short haul and long haul channel. The short haul channel is active within the
same cell site. The long haul channel is active in larger circle. The disadvantage of such dual link
is interference due to co channel problem. The second level of failure at transport layer is caused
due to link or component breakdown. This can be overcome by using mesh type of architecture at
the access layer, self healing dual rings and automated switch protection. Finally, at the top
intelligent layer, the important components constitute system databases like home location
register (HLR) and visitor location register (VLR). The survivable strategies must be
implemented in order to overcome the failures at system databases. The techniques like
distributed databases, breakpoint strategy in protocol, and backup standby databases are
recommended for robustness against failure at intelligent layer system databases. Houteo et al. [4]
analyze the reliability aspects of access network topologies ensure QoS with lowest cost for end
users. Three access network topologies are studied: the tree with parallel back up links, the ring
and the partially meshed topologies. The study establishes (i) the cost as a function of the quality
of service (ii) optimizes the partially meshed topology for more reliable networks (iii) define a
shaping policy to deal with variety of traffic schemes. To provide survivability in cellular access
network, Snow et al [4] proposed solution at architectural level like sonnet ring, devices with
multiple modes of operations and overlay networks. Sonnet Rings: Sonnet ring inhibits ring
architecture with dual rings. In case of single link cut or single transmitter/receiver failure, the
backup ring tolerates such failure as it is further connected to multiple links. Sonnet rings, target
to provide end-to-end reliability by adding redundancy. The telephone switched network is
connected with multiple base stations for same cell site area by using sonnet rings that readily
fault tolerant. Multimode devices: Snow et al further suggest devices with multiple
functionalities or with different modes of operation to tolerate the network failures, thus the
device enhances its performance by improving survivability in the wireless access network. The
survivability is achieved in multimodal devices by overlapped services provided while designing
architecture. This overlapped service ensures wireless coverage in case of link or node failure in
the access network. Thus this method impacts the performance in an effective manner, by
increasing the coverage area even during failure.Overlay network architecture: The overlay
network consists of several access points available universally to all users. The key role of
universal access point is to provide frequency assignment for requesting user, protocol
management and content adaptation. A user in an overlay network can access these access points.
In turn user is connected to wireless network depending on availability of frequency, user specific
services and type of quality required by users.
5. Conclusion
Although dependability factors like reliability, survivability, availability is taken care very well
for wired networks; there exists deficiency in wireless networks. The complexity in wireless
network is due to signal propagation, bandwidth limitation, quality of service, on demand user
request. Adding to this, the serious problem of tolerance to link and node failures within access
network exists. The various failure scenario and important survivability mechanism of wireless
access network as existing in literature has been explored in this paper. The survivability models
for like shared link, multi homing, overlay network, sonnet ring, and multimodal devices are
discussed in detail here. Table 2 provides review of five different papers based on parameters like
theoretical base, sample choice and tests. Theoretical base desires the broad base on fundaments
and whether all the important theories are considered. Sample choice depicts adequate size and
consistence within the samples. Test illustrates the suitability and supported arguments.

17
Timeliness meant for technology advancement. Limitations must be state clearly without harming
the applicability of the problem stated.
Table 2 Consolidated Survey of Research Papers
References
[1] D.Tipper, T. Dalhberg, H. Shin, “Providing Fault Tolerance in Wireless Access Networks,” IEEE
Communications Magazine, vol. 40, no.1, pp. 58-64, January 2002.
[2] P.Kubat, J. M. Smith, and C. Yum, “Design of Cellular Networks with Diversity and Capacity
Constraints,” IEEE Transaction on Reliability, vol. 49, no. 2, pp. 165-175, June 2000.
[3] U.Varshney and A. D. Malloy, “Multilevel Fault Tolerance in Infrastructure-Oriented Wireless
Networks: Framework and Performance Evaluation,” International Journal of Network Management,
vol. 16, no. 5, pp. 351-374, February 2006.
[4] F. Houeto, S. Pierre, R. Beaubrun, and Y. Lemieux, “Reliability and Cost Evaluation of Third-
Generation Wireless Access Network Topologies: A Case Study," IEEE Trans. On Reliability, vol.
51, no. 2, pp. 229-239, June 2002.
[5] A.Snow, U. Varshney, and A. Malloy, “Reliability and Survivability of Wireless and Mobile
Networks,” IEEE Computer, vol. 33, no. 7, pp. 49–55, July 2000.
[6] U.Varshney, R. Vetter, “Emerging mobile and wireless networks,” Communications of the ACM, vol.
43 no.6 pp. 73–81. 2000.

18
[7] A. Dutta, P. Kubat, “Design of partially survivable networks for cellular telecommunication systems,”
European Journal of Operational Research vol. 118, pp. 52–64, 1999.
[8] Oriol Sallen, et al., “A Roadmap from UMTS Optimization to LTE Self-Optimization”, IEEE
Communications Magazine, pp 172-182, June 2011.
[9] F. Gordejuela-Sanchez and J. Zhang, “LTE Access Network Planning and Optimization: A Service-
Oriented and Technology-Specific Perspective,” IEEE Global Communications Conference
(GLOBECOM’09), Honolulu, Hawaii, USA, December 2009.
[10] A. Quintero, S. Pierre , “On the Design of Large-Scale UMTS Mobile Networks Using Hybrid
Genetic Algorithms”, IEEE Trans On Vehicular Technology, vol. 57, no. 4. pp. 2498-2508, 2008.

A Survey on Cost Effective Survivable Network Design in Wireless Access Network

More Related Content

What's hot (15)

Similar to A Survey on Cost Effective Survivable Network Design in Wireless Access Network (20)

Recently uploaded (20)

A Survey on Cost Effective Survivable Network Design in Wireless Access Network