The Internet in particular, and large data networks (LDNW) in general, fundamentally differ from the conventional public switching telephone networks (PSTN). The occurrence of viruses, intrusions, and failures, as well as missing data and incomplete information aggravate this difference. As a result, the classic teletraffic theory – one of the most successful applications of mathematics in industry—cannot be directly applied to LDNW traffic modeling.
In this research, we propose to use renewal processes of a general type, especially their extended versions—mixtures and Markov-chain controlled renewal processes, as stochastic models for LDNW traffic.
The popular alternative approach to LDNW traffic modeling is related to fractional (self-similar) processes. This approach seems very promising in getting consistent models and results that would fit into a realistic scenario. We will develop certain self-similar stochastic models and compare them with the renewal-type models.
Stochastic Modeling of Data Networks, Identification, and Virus/Failure/Intrusion Detection
Geometric and Topological modeling of networks
Adaptive Supervisory Control Of Information Assurance Systems
Network Simulator, VINT, Packet Traces
