William Liu1, Harsha Sirisena2, Krzysztof Pawlikowski2

Slides:



Advertisements
Similar presentations
M. Dumbser 1 / 23 Analisi Numerica Università degli Studi di Trento Dipartimento dIngegneria Civile ed Ambientale Dr.-Ing. Michael Dumbser Lecture on Numerical.
Advertisements

Greening Backbone Networks Shutting Off Cables in Bundled Links Will Fisher, Martin Suchara, and Jennifer Rexford Princeton University.
Cognitive Radio Communications and Networks: Principles and Practice By A. M. Wyglinski, M. Nekovee, Y. T. Hou (Elsevier, December 2009) 1 Chapter 12 Cross-Layer.
Performance in Decentralized Filesharing Networks Theodore Hong Freenet Project.
On Scheduling of Data Dissemination in Vehicular Networks with Mesh Backhaul Liu Zhongyi M.S. Candidate, Peking Univ (To.
Fundamental Relationship between Node Density and Delay in Wireless Ad Hoc Networks with Unreliable Links Shizhen Zhao, Luoyi Fu, Xinbing Wang Department.
and 6.855J Cycle Canceling Algorithm. 2 A minimum cost flow problem , $4 20, $1 20, $2 25, $2 25, $5 20, $6 30, $
Tuesday, May 7 Integer Programming Formulations Handouts: Lecture Notes.
Graph of a Curve Continuity This curve is _____________These curves are _____________ Smoothness This curve is _____________These curves are _____________.
6.6 Analyzing Graphs of Quadratic Functions
0 - 0.
MULT. INTEGERS 1. IF THE SIGNS ARE THE SAME THE ANSWER IS POSITIVE 2. IF THE SIGNS ARE DIFFERENT THE ANSWER IS NEGATIVE.
Addition Facts
Properties of numbers EVEN and ODD numbers
Real-Time Competitive Environments: Truthful Mechanisms for Allocating a Single Processor to Sporadic Tasks Anwar Mohammadi, Nathan Fisher, and Daniel.
THERMAL-AWARE BUS-DRIVEN FLOORPLANNING PO-HSUN WU & TSUNG-YI HO Department of Computer Science and Information Engineering, National Cheng Kung University.
AMCS/CS229: Machine Learning
1 Parallel Algorithms (chap. 30, 1 st edition) Parallel: perform more than one operation at a time. PRAM model: Parallel Random Access Model. p0p0 p1p1.
1 Generating Network Topologies That Obey Power LawsPalmer/Steffan Carnegie Mellon Generating Network Topologies That Obey Power Laws Christopher R. Palmer.
Shadow Prices vs. Vickrey Prices in Multipath Routing Parthasarathy Ramanujam, Zongpeng Li and Lisa Higham University of Calgary Presented by Ajay Gopinathan.
© 2009 IBM Corporation IBM Research Xianglong Liu 1, Junfeng He 2,3, and Bo Lang 1 1 Beihang University, Beijing, China 2 Columbia University, New York,
1 Capacity analysis of mesh networks with omni or directional antennas Jun Zhang and Xiaohua Jia City University of Hong Kong.
The Capacity of Wireless Networks Danss Course, Sunday, 23/11/03.
The Capacity of Wireless Networks
Communications Research Centre (CRC) Defence R&D Canada – Ottawa 1 Properties of Mobile Tactical Radio Networks on VHF Bands Li Li & Phil Vigneron Communications.
演 算 法 實 驗 室演 算 法 實 驗 室 On the Minimum Node and Edge Searching Spanning Tree Problems Sheng-Lung Peng Department of Computer Science and Information Engineering.
Advance Mathematics Section 3.5 Objectives:
Great Theoretical Ideas in Computer Science
Quadratic Inequalities
Outcome: Determine the square root of perfect squares.
Scale Free Networks.
1 Dynamics of Real-world Networks Jure Leskovec Machine Learning Department Carnegie Mellon University
Quantitative Analysis (Statistics Week 8)
Routing and Congestion Problems in General Networks Presented by Jun Zou CAS 744.
Lecture plan Outline of DB design process Entity-relationship model
Scalable and Dynamic Quorum Systems Moni Naor & Udi Wieder The Weizmann Institute of Science.
Addition 1’s to 20.
Exponential and Logarithmic Functions
The Fundamental Theorem of Algebra
Test B, 100 Subtraction Facts
U1A L1 Examples FACTORING REVIEW EXAMPLES.
Week 1.
2 x0 0 12/13/2014 Know Your Facts!. 2 x1 2 12/13/2014 Know Your Facts!
Chapter 18: The Chi-Square Statistic
2.4 – Factoring Polynomials Tricky Trinomials The Tabletop Method.
1/22 Worst and Best-Case Coverage in Sensor Networks Seapahn Meguerdichian, Farinaz Koushanfar, Miodrag Potkonjak, and Mani Srivastava IEEE TRANSACTIONS.
Where Are You From? Confusing Location Distinction Using Virtual Multipath Camouflage Song Fang, Yao Liu Wenbo Shen, Haojin Zhu 1.
Delay Analysis and Optimality of Scheduling Policies for Multihop Wireless Networks Gagan Raj Gupta Post-Doctoral Research Associate with the Parallel.
Copyright © Cengage Learning. All rights reserved.
Distributed Constraint Satisfaction Problems M OHSEN A FSHARCHI.
13-Optimization Assoc.Prof.Dr. Ahmet Zafer Şenalp Mechanical Engineering Department Gebze Technical.
Link State Routing Jean-Yves Le Boudec Fall
5 x4. 10 x2 9 x3 10 x9 10 x4 10 x8 9 x2 9 x4.
Parallel algorithms for expression evaluation Part1. Simultaneous substitution method (SimSub) Part2. A parallel pebble game.
Computational Facility Layout
Section 3.4 The Traveling Salesperson Problem Tucker Applied Combinatorics By Aaron Desrochers and Ben Epstein.
0 x x2 0 0 x1 0 0 x3 0 1 x7 7 2 x0 0 9 x0 0.
1 Algorithmic Networks & Optimization Maastricht, November 2008 Ronald L. Westra, Department of Mathematics Maastricht University.
Count to 20. Count reliably at least 10 objects. Use ‘more’ and ‘less’ to compare two numbers. Count reliably at least 10 objects. Estimate number of objects.
7x7=.
Leting Wu Xiaowei Ying, Xintao Wu Dept. Software and Information Systems Univ. of N.C. – Charlotte Reconstruction from Randomized Graph via Low Rank Approximation.
Finding Protection Cycles in DWDM Networks 2002 IEEE ICC on Volume 5, 28 April-2 May Page(s): Reporter: Jyun-Yong Du.
The Erdös-Rényi models
A Graph-based Friend Recommendation System Using Genetic Algorithm
Spectral Analysis based on the Adjacency Matrix of Network Data Leting Wu Fall 2009.
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.
Evaluation of Geographic Vulnerabilities in Networks Todd Gardner Dr. Cory Beard.
11/02/2001 Workshop on Optical Networking 1 Design Method of Logical Topologies in WDM Network with Quality of Protection Junichi Katou Dept. of Informatics.
Random Walk for Similarity Testing in Complex Networks
Totally Disjoint Multipath Routing in Multihop Wireless Networks Sonia Waharte and Raoef Boutaba Presented by: Anthony Calce.
Presentation transcript:

A New Topological Index for Capacity Allocation Problem in Survivable Networks William Liu1, Harsha Sirisena2, Krzysztof Pawlikowski2 and Andreas Willig2 1Auckland University of Technology 2University of Canterbury New Zealand

Outline Motivation New Proposed Topological Metric : Algebraic Connectivity/Network mean distance Numerical results Future work 1 2

Motivation: Two main design issues of Survivable Networks : survivable routing scheme, and spare capacity allocation strategy. We focus on studies of quantitative structure-performance relationships and try to find the correlation between the total capacity utility and a topological index under the SBPP (Shared Backup Path Protection) scheme. Main result of this paper: Initial studies of selected network topologies confirm that a monotonically decreasing power law correlation between the total capacity and its network structure can be well described by this new topological index. 2 3

Why Metric (1/2) Most previous researches on network survivable capacity design have been generally used the average nodal degree (i.e., an arithmetic mean of set of nodal degrees) as network connectivity index In our previous DRCN paper*, we concluded that this metric is only a coarse indicator of how sparse or dense a given topology is. It carries insufficient information on network topological characteristics But our previous results are based on the similar topologies, and do not consider the different sizes and shapes of the topologies How to find a proper topological metric to describe different shapes and sizes of topologies? *William Liu, Harsha Sirisena, Krzysztof Pawlikowski and Allan McInnes, Utility of Algebraic Connectivity Metric in Topology Design of Survivable Networks, 7th International Workshop on the Design of Reliable Communication Networks (DRCN 2009), Washington DC, October 25-28, 2009. 3 4

Why Metric (2/2) Therefore… There is an obvious need for a more informative metric that can capture the topological characteristic of a network : size and shape Algebraic connectivity PLUS Network mean distance introduced in spectral graph theory, are key topological invariants of graphs, which can indicate the connectivity, size and shape of a topology 4 5

Algebraic Connectivity Definition Given an unweighted graph G with m links and n nodes, with its adjacency matrix Anxn and degree matrix Dnxn Matrix Lnxn= D-A is known as the Laplacian matrix of G Spectrum of graph G = the set of n eigenvalues of matrix L The algebraic connectivity of G: λ2(G)=the 2nd smallest eigenvalue of Lnxn , also called the Fiedler value in spectral graph theory. It provides global information about graph’s topological properties: it is larger for stronger connected graphs. 5

Example- Adjacency Matrix Adjacency matrix (A) a square matrix n x n, where Aij is 1 if vertex xi and xj is connected, otherwise is 0 x1 x2 x3 x4 x5 x6 1 5 1 6 2 4 3 Important properties: Symmetric matrix Its eigenvectors are real and orthogonal 6

Example- Degree Matrix Degree matrix (D) n x n diagonal matrix where Dii is the nodal degree of vertex xi x1 x2 x3 x4 x5 x6 3 2 5 1 6 2 4 3 7

Example- Laplacian Matrix Laplacian matrix (L) n x n symmetric matrix L=D-A x1 x2 x3 x4 x5 x6 3 -1 2 5 1 6 2 4 3 Important properties: Eigenvalues are non-negative real numbers Eigenvectors are real and orthogonal Eigenvalues and eigenvectors provide an insight into the connectivity of the graph 8

Properties of The algebraic connectivity is a robust measure of network connectivity: the larger the algebraic connectivity, the better connected the network. Hence more node- and link-disjoint paths exist between pairs of nodes. In addition, the network mean distance here is defined as the average of all hop count between distinct nodes of the graph. This metric can measure the size and shape of the graph. 9

Properties of The bounds on the network mean distance can be derived*. Its lower bound is: and its upper bound is: We propose to use to reflect network topological properties such as the connectivity, size and shape of a given topology and it has impacts on the total capacity allocation in network survivability design, * Bojan Mohar, “Eigenvalues, Diameter and Mean Distance in graphs,” Graphs Combin. (1991), pp. 53–64. 10

Simulation Scenario Uniform traffic Shared Backup Path Protection (SBPP) AMPL/CPLEX optimization tool Topologies: 10 different topologies Performance: Total capacity, i.e., working capacity + spare capacity Curve-fitting : the coefficient of determination R2 (0,1), with a higher value indicating a better fit. 11

Topologies 12

Experiment results

Experiment results Total capacities vs. four topological indices 14

Experiment results Total capacity vs. average node degree 15

Experiment results Total capacity vs. algebraic connectivity/exact mean distance * The exact mean distance can be calculated by the standard graph toolbox in MATLAB

Experiment results Total capacity vs. algebraic connectivity/lower bound mean distance

Experiment results Total capacity vs. algebraic connectivity/upper bound mean distance

Conclusions and Future Work Our initial results show that for the chosen example networks, the new metric is a good quantitative descriptor of network topologies in studies of the capacity allocation problem. We have found a power law relationship between this topological index and total capacity in survivable network designed with the SBPP algorithm. This new finding can be applied into network optimization, for such as the estimation on the total capacity if the topology is given, or on the other hand, the evaluation on the topology structure if the total capacity constraints are given. Future works… An extensive validation on our findings on more complex topologies is underway. Further studies on how to apply this new topological index into network capacity planning& optimization, e.g., one node/link addition problem are needed. 19

Questions Please address your questions to william.liu@aut.ac.nz