Presentation is loading. Please wait.

Presentation is loading. Please wait.

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

Similar presentations


Presentation on theme: "Modeling Malware Spreading Dynamics Michele Garetto (Politecnico di Torino – Italy) Weibo Gong (University of Massachusetts – Amherst – MA) Don Towsley."— Presentation transcript:

1 Modeling Malware Spreading Dynamics Michele Garetto (Politecnico di Torino – Italy) Weibo Gong (University of Massachusetts – Amherst – MA) Don Towsley (University of Massachusetts – Amherst – MA) INFOCOM 2003

2 2 Outline Motivation Modeling framework Interactive Markov Chains Analysis and simulation of malware propagation dynamics Percolation problem Transient behavior Conclusions

3 3 Motivation The Internet is an easy and powerful mechanism for propagating malicious software programs : the “ malware ” ( viruses, worms, …) It is expected that future malware acitivity will be more prevalent and virulent, resulting in significant greater damage and economic losses

4 4 Motivation Dynamics of malware propagation are still not well understood We would like to:  Predict the temporal evolution of an infection process that starts propagating on a network  Design and evaluate effective countermeasures  Assess the defensibility and vulnerability of different network architectures Need to develop mathematical methodologies that are able capture the spreading characteristics of malware

5 5 Our contribution A flexible modeling framework based on Interactive Markov Chains, able to capture the probabilistic nature of malware propagation Application of such framework to the case of viruses: Identification of a “percolation problem” Investigation of the impact of the underlying network topology Analytical bounds and approximations validated through extensive simulations

6 6 The “Interactive Markov Chain” (IMC) Modeling Framework  Global network structure... alert failednormal Local Structure (can vary from node to node) secureinsecureemergency  Each node is represented by a Markov chain, whose state transitions are influenced by the status of its neighbors but locally a Markov chain Global Structure (the network)

7 7 Computational complexity issue  The whole system evolves according to a global Markov chain G, whose state space dimension ( #G ) is equal to the product of the local chain dimensions ( #L )  G =  L N  The solution of the global Markov chain is feasible only for small systems example: - 20 nodes - binary status (0 = not infected, 1 = infected)  2 20 states !

8 8 How can we study very large systems (thousands of nodes) ?  Discrete event simulations of the model  how many runs ? how long ?  could be computationally too expensive  do not help to understand the system dynamics  Analytical bounds and approximations  quick prediction of the system behavior  gross-level approximations can be sufficient  provide insights into the inner dynamics

9 9 The virus propagates as attachment to messages Requires human assistance  random time elapses before the recipient reads the message  the “click” probability  The virus makes use of the recipient’s address book to send copies of itself virus propagation

10 10 We consider the network graph induced by address books Each node stands for an address Edges represent social or business relationships The resulting graph is expected to have “small world” properties:  small characteristic path length  high clustering coefficient IMC model : global structure

11 11 IMC model : local structure 3 statuses for each node: S (Susceptible): the node can be infected by the virus I (Infected): the node has been infected by the virus M (Immune): the node can no more be infected by the virus Discrete time model probability that node “j” is susceptible at time k probability that node “j” is infected at time k probability that node “j” is immune at time k

12 12 IMC model i j w ij IM S cjcj 1-c j

13 13 Virus propagation model  The numerical solution of the system requires to know the joint probabilities of neighboring nodes: ?

14 14 Fundamental questions about malware propagation dynamics: Virus propagation model  What is the final size of the infection outbreak ?  How many nodes (on average) will be reached by the virus at the end ?  How fast is the malware propagation ?  What is the (average) number of infected node as a function of time ?

15 15 The “small-world” model of Watts and Strogatz A ring lattice with additional random shortcuts Parameters: N = number of nodes S = number of shortcuts (  = shortcuts density) k = lattice connectivity (number of neighbors on each side of a node) N = 24 S = 4 k = 3

16 16 What is the final size of the infection outbreak ? Not all of the susceptible nodes necessarily receive a copy of the virus !  site percolation problem (node occupation probability = click probability)

17 17 Site percolation problem on the small world graph: exact asymptotic result [Moore, Newman 2000]

18 18 Transient analysis of an infection process : Bounds Probability that a node has been reached by the virus Lower bound property of joint probabilities Upper bound theory of Associated Random Variables

19 19 Analytical bounds Average number of infected nodes time sim upper bound lower bound  Infinite unidimentional lattice k = 1 k = 10 k = 100 (click probability = 1) 1

20 20 Transient analysis of an infection process : approximation  Linear mixing of lower bound and upper bound k = local connectivity of the node s = self-influence probability

21 21 Fitting of mixing coefficient M(k,s) Infinite unidimentional lattice

22 22 Approximate analysis on the small world graph: the impact of topology Average number of infected nodes time model approx simulation k = 10 no shortcuts k = 10 2 shortcuts k ~ geom(10) no shortcuts k = shortcuts Fully-connected graph (2000 nodes - click probability = 1)

23 23 Combining transient analysis and percolation on general topologies  Probability not to be reached by the virus = initial immunization  overestimate of the spreading rate of the virus  Upper bound of the reaching probability on general topologies:  Global upper bound for the infection process

24 Average number of infected nodes time simulation model approx + bound perc Transient analysis and percolation on power-law random graphs nodes GLP algorithm (Bu 2002) power-law node degree, small-world properties (click probability = 0.5) m = initial connectivity m = 1 m = 2 m = 4 One initially infected node with degree 10

25 25 Conclusions We have proposed an analytical framework to study the dynamics of malware propagation on a network We have obtained useful bounds and approximations to study an infection process on a general topology Approach suitable to analyze a wide range of “dynamic interactions on networks” (routing protocols, p2p,…)

26 26 The End Thanks…

27 27 Site percolation on a given small-world graph Reaching probability node index sim upper bound - h = 0 upper bound - h = 8 approximation lower bound - h = 8 lower bound - h = 0


Download ppt "Modeling Malware Spreading Dynamics Michele Garetto (Politecnico di Torino – Italy) Weibo Gong (University of Massachusetts – Amherst – MA) Don Towsley."

Similar presentations


Ads by Google