Title: Fault Tolerant Deployment and Topology Control in Wireless Networks
1Fault Tolerant Deployment and Topology Control
in Wireless Networks
MOBIHOC 2003
X.-Y. Li, P.-J. Wan, Y. Wang, and C.-W. Yi Dept.
of Computer Science Illinois Institute of
Technology
- Presented by Ning Li
- November 22, 2009
2Outline
- Introduction
- Related Work
- Fault Tolerant Deployment
- Topology Control for Fault Tolerance
- Simulations
3Introduction
- k-connected (k-vertex connect) if for each pair
of vertices, there are k mutually vertex disjoint
(except end-vertices) paths connecting them.
Equivalently, a graph is k-connected if there is
no a set of k-1 nodes whose removal will
partition the network into at least two
components. - k-edge connected if for each pair of vertices,
there are k mutually edge disjoint paths
connecting them. - Fault Tolerance a k-connected wireless network
can sustain the failure of k-1 nodes.
4Point Process
- Uniform Random Point Process Xn consists of n
independent points each of which is uniformly and
independently distributed over a region ?. - Homogeneous Poisson Process the number of points
in a region is a random variable depending only
on the volume of the region
5Related Work (1)
- P. Gupta and P. R. Kumar. Critical power for
asymptotic connectivity in wireless networks.
Stochastic Analysis, Control, Optimization and
Applications, 1998. - Assumptions
- n nodes are independently and uniformly
distributed in a disk of unit area in ?2 - Assume a common range rn
- Two nodes can communicate with each other if the
distance is ? rn. - Objective Find the critical range rn.
- Results Let , the network is
asymptotically connected w.p.1 if and is
asymptotically disconnected w.p.1 if
6Related Work (2)
- M. Penrose. The longest edge of the random
minimal spanning tree. Annals of Applied
Probability, 1997. - Results The longest edge Mn of the minimum
spanning tree of n points randomly and uniformly
distributed in a unit area square C satisfies
that
7Related Work (3)
- F. Xue and P. R. Kumar, The number of neighbors
needed for connectivity of wireless networks. To
appear in Wireless Networks. - Assumptions
- n nodes are independently and uniformly
distributed in a square of unit area in ?2 - Each node is connected to its ?n nearest
neighbors. - Objective Find ?n such that the resulting
network topology is asymptotically connected. - Results ?(log n) neighbors are necessary and
sufficient, i.e., there are two constants such
that (c10.074 and
c2gt2/log(4/e)5.1774)
8Related Work (4)
- M. Penrose. On k-connectivity for a geometric
random graph. Random Structures and Algorithms,
1999. - Assumptions
- n nodes are independently and uniformly
distributed in a unit cube in ?d - Two nodes can communicate with each other if the
distance is ? r (common range). - Results Let ?n denote the minimum r at which the
graph is k-connected and let ?n denote the
minimum r at which the graph has minimum degree
k, then P?n?n?1 as n??.
9Fault Tolerant Deployment
- How is the transmission range related to the
number of nodes in a fixed area such that the
resulted network can sustain k fault nodes with
high probability? - Consider a homogeneous Poisson point process of
rate n, denoted by Pn, on a unit-area square C.
10Preliminaries
- Let denote the expected number of points of
Pn with degree k in a graph of G(Pn, r). - Let D(x,r) be the disk centered at x with radius
r. - Given a point x, let vr(x) be the area of the
intersection of D(x,r) with the unit-area square
C. - The probability that point x has degree k is
- We have
11Preliminaries
Where is the minimum radius r at which
the graph G(Pn, r) has the minimum degree at
least k1.
12(No Transcript)
13Lower Bound
is a monotone increasing function of ,
14Upper Bound
15Upper Bound
- Let w(x) be any function s.t. w(x) vr(x) and is
monotone increasing of r. - Let
- Thus,
- Let r be the solution of
- Let r be the solution of
- r r.
16Upper Bound
17Topology Control for Fault Tolerance
- Given a k-fault-tolerant deployment of wireless
nodes, find a localized method to control the
network topology such that the resulting topology
is still fault tolerant but with much fewer
communication links maintained. - We show that the constructed topology has only
linear number of links and is a length spanner. - Let ?G(u,v) be the shortest path connecting u and
v in a weighted graph G, and ?G(u,v) be the
length of ?G(u,v). Then a graph G is a t-spanner
of a graph H if V(G)V(H) and, for any two nodes
u and v of V(H), ?G(u,v)t?H(u,v). We also
call t the length stretch factor of the spanner G.
18Yao Graph
- Yao Graph at each node u, any p equal-separated
rays originated at u define p equal cones. In
each cone, choose the shortest (directed) edge uv
? G, if there is any, and add a directed link uv.
Ties are broken arbitrarily. - For fault tolerance in each cone, node u chooses
the k1 closest nodes in that cone, if there is
any, and add directed links from u to these nodes.
19Yp,k1 is a spanner
Proof Need to show that Let We have Stretch
ratio
20Simulations Probability of k-connected
21Probability of a graph with minimal degree k is
k-connected
22(No Transcript)
23(No Transcript)
24(No Transcript)