Spatial-temporal modeling of malware propagation in networks.

Area of Science:

• Computer Science
• Network Security
• Machine Learning

Background:

• Malware propagation poses a significant threat to network security and management.
• Topological scanning malware utilizes network topology for spreading, necessitating specialized models.
• Understanding malware spread is crucial for damage assessment and developing effective countermeasures.

Purpose of the Study:

• To develop and evaluate models for the spread of topological malware.
• To approximate complex spatial-temporal malware propagation processes using simpler models.
• To enhance network security strategies by accurately predicting malware behavior.

Main Methods:

• Utilized probabilistic graphs and spatial-temporal random processes to model malware spread.
• Proposed the independent model and the Markov model as approximations for spatial dependence.
• Conducted theoretical analysis and extensive simulations on diverse network topologies.

Main Results:

• The independent model effectively captures temporal dependence and network topology, surpassing existing models.
• The Markov model demonstrates improved accuracy by incorporating spatial dependence for transient and equilibrium behaviors.
• Both models were validated through theoretical analysis and simulations on large-scale networks.

Conclusions:

• The independent model offers a robust approach for understanding temporal aspects of malware propagation.
• The Markov model provides enhanced accuracy for characterizing both short-term and long-term malware spread dynamics.
• These models contribute to improved network security by offering better insights into topological malware behavior.