Adviser: Frank,Yeong-Sung Lin Present by Limin Zheng Gunhak Lee, Alan T. Murray
40
Maximal covering with network survivability requirements in wireless mesh networks Adviser: Frank,Yeong-Sung Lin Present by Limin Zheng Gunhak Lee , Alan T. Murray
Adviser: Frank,Yeong-Sung Lin Present by Limin Zheng Gunhak Lee, Alan T. Murray
Introduction Many US cities and Countries are attempting to
build wireless broadband networks for communication and service in
their communities as basic infrastructure to facilitate local
economic development and enable much wider service provision to
more people. wireless networks in municipalities have been widely
utilized for a range of public applications, such as public
hotspots, public safety and general communication Wireless
broadband networks would play an important role in improving the
quality of our life, giving people the freedom and capability to
communicate with the world anytime, anywhere Advanced wireless
broadband technologies, such as Wi-Fi, WiMax and cellular systems,
relying on mesh or multi-hop networking. Wireless broadband is
attractive to municipalities willing to construct their own
communication network given limited budgets.
Slide 5
Introduction What is the primary concern in providing wireless
broadband services ? When local governments attempt to provide
wireless broadband services to their communities, the primary
concern is where to place relevant facilities and how to connect
them. What is the purpose of this paper ? In this paper, we address
location modeling approaches for integrating maximal covering and
survivable network design in planning citywide wireless broadband
services. More specifically, we propose a mathematical formulation
of the maximal covering problem with survivability constraints
based on wireless mesh network topology.
Slide 6
Backgroud
Slide 7
Background Survivable network design What is survivable network
? What is disjoint path?
Slide 8
Background Wi-Fi based mesh networks What is mesh networks
Slide 9
Problem Description
Slide 10
For the purpose of this paper, we specifically address two
issues: (1)how to locate Wi-Fi equipment to maximally cover demand
given a specified number of units. (2)how to connect Wi-Fi
equipment to ensure survivable networking. Solution: For(1)Maximal
Covering Location Problem (MCLP) For(2) Number of node disjoint
paths for any pair of nodes
Slide 11
Problem Description Regarding the architecture of a mesh
wireless network, some of nodes (gateways) must be connected to
hard, land based infrastructure and thus reliable performance of
network is dependent on the existence of duplicated paths between a
general node and gateway node.
Slide 12
Mathematical Formulation
Slide 13
Model: Maximal Covering problem with Survivability Constraints
(MCSC) Assumed based on equipment capabilities : 1.Maximum distance
of wireless access form the Wi-Fi router. 2.Maximum distance for
wired access from the exiting backbone infrastructure. 3.Maximum
distance for interconnection between the Wi-Fi routers.
Slide 14
Mathematical Formulation Something need to predefined or given
when using this model : 1.Potentially eligible sites to provide
wireless broadband services constitute a discrete set of locations.
2.Set of points is predefined to represent aggregate population to
be covered by the Wi-Fi router. 3.A number of facilities, p, is
given exogenously. 4.a number of Wi-Fi routers, q, providing wired
connection to the existing backbone infrastructure, is also
specified in advance.
Slide 15
Mathematical Formulation Based on the hierarchy of a wireless
broadband network, parameters and sets are defined as follows: Iset
of demand nodes Jset of potential sites for Wi-Fi router Mset of
existing DSL central offices aiai population at demand node I
prequired number of Wi-Fi routers to be deployed qrequired number
of Wi-Fi routers for wired connections to existing central offices
Krequired number of disjoint paths d ij shortest distance from
demand node i to Wi-Fi router at j d jc shortest distance from
Wi-Fi router at j to DSL central office at c d jl shortest distance
between Wi-Fi routers at j and l
Slide 16
Mathematical Formulation NiNi {j J|d ij R} {j J|d cj L, c M} jj
{l J|d jl W, j l} Rcoverage standard for Wi-Fi; Lcoverage standard
for DSL central office Wmaximum distance for Wi-Fi router point to
point interconnection
Slide 17
Mathematical Formulation Decision variables are defined as
follows:
Slide 18
Mathematical Formulation
Slide 19
Slide 20
Slide 21
Application Details
Slide 22
1.It is assumed that wireless routers must be within 12,000
feet (L) from a central office. 2.The coverage standard of a Wi-Fi
router (R) is specified as 3465 feet, so 6930 feet is used for the
maximum distance for Wi-Fi point to point interconnection (W) 3.The
required number of Wi-Fi routers (p) is specified to be in the
range of 829. 4.It is assumed that 20% of Wi-Fi routers satisfy the
required number of wired connections to existing central offices
(q). 5.Two cases of disjoint paths, K = 1 and K = 2, are examined
for survivable network design.
Slide 23
Slide 24
Application Details The MCSC was solved exactly using a
commercial optimization solver, named CPLEX 10.0 (ILOG) on an Intel
Xeon 3 GHz CPU with 3 GB memory. ArcGIS 9.1 was used to manage
needed input information (Ni, W, xj) through spatial analysis
functionality. Also, Visual Basic Application (VBA) with ArcObjects
was used to create the necessary text file of the MCSC that is read
into CPLEX. Further, GIS provides capabilities for visualizing and
evaluating solutions.
Slide 25
Results & Discussion
Slide 26
Slide 27
Slide 28
Slide 29
Slide 30
Slide 31
Discussion In this paper, we focus on system reliability for
comparison between two different network configurations. For each
network configuration, there are specific source and destination
nodes. Accordingly, sets are defined as follows:
Slide 32
Evaluating system reliability Reliability of a node can be
defined as the probability that it functions during a specified
time period. Given node reliability, the probability of a disjoint
path for a pair of source and destination nodes can be derived by
the joint probability of nodes along the disjoint path, based upon
the assumption of independence. Since there could be a number of
disjoint paths between source and destination nodes, system
reliability is the sum of the probabilities of all possible
disjoint paths between source and destination nodes.
Slide 33
Evaluating system reliability The standard mathematical
formulation of system reliability can be found in Shier (1991), and
stated as follows:
Slide 34
Evaluating system reliability Average system reliability for
entire network, Raverage, is computed by averaging the
reliabilities for all pairs of source and destination nodes as
follows: where Q is the number of all pairs of source and
destination nodes.
Slide 35
Evaluating system reliability Node reliability probabilities
are assumed (0.8 in our case) Kp(s)q(t) S-t pairs (Q) Disjoint path
R average 111660710.69 211440800.41 For examining these two
configurations in cases of a specific node failure, we calculate
reliability after simulating any single node failure. K=1, average
reliability = 0.5 K=2, average reliability = 0.61
Slide 36
Conclusions
Slide 37
Adequately positioning wireless access points is crucial in
order to extend service coverage with a given budget limit. Another
significant consideration for building wireless broadband networks
is the provision of reliable broadband service. However, it is
difficult to cover a large area reliably because a more reliable
broadband network ecessarily requires a more interconnected network
topology to ensure redundancy in routing.
Slide 38
Conclusions To deal with these considerations simultaneously,
we introduced the maximal covering problem with survivability
requirements (MCSC). This approach extends classical facility
location and network design problems by explicitly integrating
covering and network survivability. For more practical use of this
approach, several related technical issues, such as radio coverage
planning, traffic and routing controls and channel assignment, must
be taken into account. This paper, however, focuses on general
methodological issues concerning maximal covering and survivable
network design. Thus, this paper is expected to help decision
makers and network planners understand coverage and design issues
through the use of a method for obtaining solutions and presenting
expected network configurations.
Slide 39
Conclusions The application found that a wireless network can
be designed to provide citywide wireless broadband services to an
urban area, ensuring network survivability to a high degree.
Comparatively, we also highlighted two types of tradeoffs. One
tradeoff exits between coverage and the level of survivability.
Another one exits between coverage and total cost.