Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland CONTROLLING EPIDEMICS IN WIRELESS NETWORKS Ranjan.

Slides:



Advertisements
Similar presentations
Modeling Malware Spreading Dynamics Michele Garetto (Politecnico di Torino – Italy) Weibo Gong (University of Massachusetts – Amherst – MA) Don Towsley.
Advertisements

It’s a Small World by Jamie Luo. Introduction Small World Networks and their place in Network Theory An application of a 1D small world network to model.
Tirgul 7 Review of graphs Graph algorithms: –DFS –Properties of DFS –Topological sort.
Beyond Trilateration: On the Localizability of Wireless Ad Hoc Networks Reported by: 莫斌.
 Graph Graph  Types of Graphs Types of Graphs  Data Structures to Store Graphs Data Structures to Store Graphs  Graph Definitions Graph Definitions.
1 Epidemic Spreading in Real Networks: an Eigenvalue Viewpoint Yang Wang Deepayan Chakrabarti Chenxi Wang Christos Faloutsos.
Graph & BFS.
Beneficial Caching in Mobile Ad Hoc Networks Bin Tang, Samir Das, Himanshu Gupta Computer Science Department Stony Brook University.
Graph COMP171 Fall Graph / Slide 2 Graphs * Extremely useful tool in modeling problems * Consist of: n Vertices n Edges D E A C F B Vertex Edge.
CSE 780 Algorithms Advanced Algorithms Graph Algorithms Representations BFS.
Shortest path algorithm. Introduction 4 The graphs we have seen so far have edges that are unweighted. 4 Many graph situations involve weighted edges.
Online Data Gathering for Maximizing Network Lifetime in Sensor Networks IEEE transactions on Mobile Computing Weifa Liang, YuZhen Liu.
Graphs G = (V,E) V is the vertex set. Vertices are also called nodes and points. E is the edge set. Each edge connects two different vertices. Edges are.
The Shortest Path Problem
STOCHASTIC GEOMETRY AND RANDOM GRAPHS FOR THE ANALYSIS AND DESIGN OF WIRELESS NETWORKS Haenggi et al EE 360 : 19 th February 2014.
1 The Orphan Problem in ZigBee- based Wireless Sensor Networks IEEE Trans. on Mobile Computing (also in MSWiM 2007) Meng-Shiuan Pan and Yu-Chee Tseng Department.
A Node-Centric Load Balancing Algorithm for Wireless Sensor Networks Hui Dai, Richar Han Department of Computer Science University of Colorado at Boulder.
1 Topology Control of Multihop Wireless Networks Using Transmit Power Adjustment Infocom /12/20.
Introduction to Graph Theory
The Erdös-Rényi models
Presented by: Chaitanya K. Sambhara Paper by: Maarten Ditzel, Caspar Lageweg, Johan Janssen, Arne Theil TNO Defence, Security and Safety, The Hague, The.
Fundamental Lower Bound for Node Buffer Size in Intermittently Connected Wireless Networks Yuanzhong Xu, Xinbing Wang Shanghai Jiao Tong University, China.
Developing Analytical Framework to Measure Robustness of Peer-to-Peer Networks Niloy Ganguly.
Computing and Communicating Functions over Sensor Networks A.Giridhar and P. R. Kumar Presented by Srikanth Hariharan.
GRAPHS CSE, POSTECH. Chapter 16 covers the following topics Graph terminology: vertex, edge, adjacent, incident, degree, cycle, path, connected component,
A Clustering Algorithm based on Graph Connectivity Balakrishna Thiagarajan Computer Science and Engineering State University of New York at Buffalo.
GRAPH SPANNERS by S.Nithya. Spanner Definition- Informal A geometric spanner network for a set of points is a graph G in which each pair of vertices is.
Lecture7 Topic1: Graph spectral analysis/Graph spectral clustering and its application to metabolic networks Topic 2: Different centrality measures of.
Data Structures Week 9 Introduction to Graphs Consider the following problem. A river with an island and bridges. The problem is to see if there is a way.
Efficient Deployment Algorithms for Prolonging Network Lifetime and Ensuring Coverage in Wireless Sensor Networks Yong-hwan Kim Korea.
1 Multicast Algorithms for Multi- Channel Wireless Mesh Networks Guokai Zeng, Bo Wang, Yong Ding, Li Xiao, Matt Mutka Michigan State University ICNP 2007.
ENERGY-EFFICIENT FORWARDING STRATEGIES FOR GEOGRAPHIC ROUTING in LOSSY WIRELESS SENSOR NETWORKS Presented by Prasad D. Karnik.
Graphs. Definitions A graph is two sets. A graph is two sets. –A set of nodes or vertices V –A set of edges E Edges connect nodes. Edges connect nodes.
Xiaowei Ying, Xintao Wu Dept. Software and Information Systems Univ. of N.C. – Charlotte 2008 SIAM Conference on Data Mining, April 25 th Atlanta, Georgia.
Analyzing the Vulnerability of Superpeer Networks Against Attack Niloy Ganguly Department of Computer Science & Engineering Indian Institute of Technology,
Synchronization in complex network topologies
By Naeem Amjad 1.  Challenges  Introduction  Motivation  First Order Radio Model  Proposed Scheme  Simulations And Results  Conclusion 2.
Efficient Computing k-Coverage Paths in Multihop Wireless Sensor Networks XuFei Mao, ShaoJie Tang, and Xiang-Yang Li Dept. of Computer Science, Illinois.
Covering Points of Interest with Mobile Sensors Milan Erdelj, Tahiry Razafindralambo and David Simplot-Ryl INRIA Lille - Nord Europe IEEE Transactions on.
KAIS T On the problem of placing Mobility Anchor Points in Wireless Mesh Networks Lei Wu & Bjorn Lanfeldt, Wireless Mesh Community Networks Workshop, 2006.
Performance of Adaptive Beam Nulling in Multihop Ad Hoc Networks Under Jamming Suman Bhunia, Vahid Behzadan, Paulo Alexandre Regis, Shamik Sengupta.
Domain decomposition in parallel computing Ashok Srinivasan Florida State University.
Data Structures & Algorithms Graphs Richard Newman based on book by R. Sedgewick and slides by S. Sahni.
Graphs G = (V,E) V is the vertex set. Vertices are also called nodes and points. E is the edge set. Each edge connects two different vertices. Edges are.
Brief Announcement : Measuring Robustness of Superpeer Topologies Niloy Ganguly Department of Computer Science & Engineering Indian Institute of Technology,
CS 590 Term Project Epidemic model on Facebook
1 11 Distributed Channel Assignment in Multi-Radio Mesh Networks Bong-Jun Ko, Vishal Misra, Jitendra Padhye and Dan Rubenstein Columbia University.
1 Finding Spread Blockers in Dynamic Networks (SNAKDD08)Habiba, Yintao Yu, Tanya Y., Berger-Wolf, Jared Saia Speaker: Hsu, Yu-wen Advisor: Dr. Koh, Jia-Ling.
Load Balanced Link Reversal Routing in Mobile Wireless Ad Hoc Networks Nabhendra Bisnik, Alhussein Abouzeid ECSE Department RPI Costas Busch CSCI Department.
Chapter 9: Graphs.
1 Low Latency Multimedia Broadcast in Multi-Rate Wireless Meshes Chun Tung Chou, Archan Misra Proc. 1st IEEE Workshop on Wireless Mesh Networks (WIMESH),
Indian Institute of Technology Kharagpur PALLAB DASGUPTA Graph Theory: Trees Pallab Dasgupta, Professor, Dept. of Computer Sc. and Engineering, IIT
Design and Analysis of Algorithms Introduction to graphs, representations of a graph Haidong Xue Summer 2012, at GSU.
Xiaowei Ying, Kai Pan, Xintao Wu, Ling Guo Univ. of North Carolina at Charlotte SNA-KDD June 28, 2009, Paris, France Comparisons of Randomization and K-degree.
Trees.
Random Walk for Similarity Testing in Complex Networks
BIPARTITE GRAPHS AND ITS APPLICATIONS
Analysis of Node Localizability in Wireless Ad-hoc Networks
CSC317 Graph algorithms Why bother?
Network analysis.
Network Science: A Short Introduction i3 Workshop
Graphs All tree structures are hierarchical. This means that each node can only have one parent node. Trees can be used to store data which has a definite.
Graphs Chapter 11 Objectives Upon completion you will be able to:
Graph Operations And Representation
Trees.
Chapter 11 Graphs.
Barrier Coverage with Optimized Quality for Wireless Sensor Networks
Text Book: Introduction to algorithms By C L R S
Graphs G = (V,E) V is the vertex set.
INTRODUCTION A graph G=(V,E) consists of a finite non empty set of vertices V , and a finite set of edges E which connect pairs of vertices .
Presentation transcript:

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland CONTROLLING EPIDEMICS IN WIRELESS NETWORKS Ranjan Pal 1 Ayan Nandy 2 Satya Ardhy Wardana 3 Neeli Rashmi Prasad 3 Ramjee Prasad 3 1 University of Southern California, USA 2 Indian Institute of Technology, Kharagpur, India 3 Center for TeleInfrastruktur (CTIF), Aalborg University, Denmark

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland 2 OUTLINE Introduction Probem Definition Network Model Strategy Selection Algorithm Conclusion

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland INTRODUCTION Modern day wireless communications is witnessing the emergence of various viruses that can spread over air interfaces. Several application scenarios in multi-hop wireless networks require many nodes to cooperate on a single application.  Wireless Sensor Network  Social Network Viruses can spread over the air quickly from device to device and devastate the entire network in a short period of time  causing an EPIDEMIC 3

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland PROBLEM DEFINITION We propose a novel method using spectral properties of graphs to select the best suited epidemic control strategy for a wireless network from a pool of strategies. Our methodology will provide a general framework for good, structured, centralized strategy selection for epidemic control in static multihop wireless networks. 4

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland NETWORK MODEL Multi-hop network Undirected Single channel, single radio per node Network represented by a graph G = (V,E) Links (u,v) and (v,u) are different V represents the set of nodes and E is the set of wireless data links 5 uv 5

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland EXAMPLE OF LAPLACIAN MATRIX OF A GRAPH 6 Spectral function of the network graph the eigenvalues of the Laplacian of a graph  Centrality of its vertices

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland NETWORK MODEL We term this graph as a parent graph. Spread Graph - Initial configuration : a group of independent nodes in 2-D space. - a compromised node spreads its virus to its neighboring nodes after k time units of it being infected.  to model the spreading of the virus (evolution) - Monitor the spread graph at certain pre-specified time intervals, then apply our control strategy to the graph at each interval. We assume that it is not possible to monitor the network all the time 7

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland STRATEGY SELECTION ALGORITHM The spectrum of a parent graph G is the set of eigenvalues of its Laplacian matrix L. A spectral function F of a parent graph G is formed, taking as arguments of the eigenvalues of its Laplacian. We define our spectral function as the average of all the igenvalues. The main aim is to be able to decide the best epidemic control strategy from a pool of strategies for a given parent graph. We classify parent subgraphs into 2 categories (predefined threshold): 1) graphs with low values of F (below the threshold) 2) graphs with high values of F (above the threshold) 8

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland STRATEGY SELECTION ALGORITHM We observe its spread graph over time and apply each of our proposed strategies at certain pre-specified time intervals. Doing so over, a total inspection period of T will result in ranking the strategies in order of effectiveness of epidemic check. We then try to see whether there is any strategy that suits best for all the parent graphs in the set with high values of F. Likewise, we try to find strategies which suit best for all parent graphs in the set with low values of F. 9

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland 10 CONCLUSION We have proposed a spectral graph-theoretic selection algorithm that aims at selecting the best strategy from a pool of strategies, to prevent an epidemic outbreak in a given parent network. As part of ongoing work, we are performing a detailed simulation study based on our algorithm to come up with a general strategy classification scheme.

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland 11 THANK YOU

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland STRATEGIES ALGORITHM Strategy 1: Treatment of edges whose end vertices result in maximum sum of degrees. Strategy 2: Treatment of edges whose end vertices result in maximum product of degrees. Strategy 3: Treatment of vertices whose degree is the Maximum. Strategy 4: Treatment of edges whose end vertices result in maximum harmonic product of degrees 12

Workshop on Applications of Wireless Communications (WAWC 2008) 21 August 2008, Lappeenranta - Finland STRATEGIES ALGORITHM A. Functional Properties One way to comment on the spread of virus in an evolving spread network is to observe the values of its functional properties over time. The most common functional property used is characteristic path length [5]. In this paper we use a different functional property called harmonic path length, to track the spread of virus. We define harmonic path length (HPL) of any graph G as the median of the harmonic means of the shortest path lengths connecting each vertex in G to all other vertices. If a vertex in a graph is isolated, it’s distance from any other vertex in the graph is infinite. The characteristic path length (CPL) with respect to a vertex is the average of the shortest distance from that vertex to every other vertex in the graph. The CPL of the graph is the median of CPL value over all the vertices. Now, even if one vertex is isolated, the CPL for each vertex will be infinite. So, the CPL of the graph will be infinite. In this paper, we will be mostly dealing with ’evolving’ spread graphs which are generally sparse/disconnected in nature. To compare between two different graphs, each having disjoint subgraphs, we need a different parameter. That is why we use harmonic path length which comes up with a finite value if at least half the vertices of the graph are not isolated, else the value is infinite. More than one graph might have an infinite value. To distinguish between them, the time when infinity is reached for each, is an important parameter. The next subsection elaborates this point. B. Suitability of Strategies Functional properties determine the suitability of a particular strategy. For a particular strategy, an infinite value of HPL at a time instant indicates that more than half of a spread network is isolated. This means that the virus has not been able to substantially percolate in the network. Let Gst be the spread graph under strategy s at time instant t. If the value of HPL of Gst at t is infinity and remains so for the interval [t; T], where T is a large enough total inspection period, the strategy s under question is said to be successful in epidemic check. Lesser the value of t, better is the strategy. If t > T then the strategy under question does not cause the cessation of virus spread and is considered not that effective. In Figure 1, we show a typical variation of HPL with time in an example spread network. The value 200 on the y-axis represents infinity. The spread of virus is in full effect between time instants 35 and 175, after which there begins an improvement in the situation. The virus spread decreases considerably after time instant