| Various anonymizing services and networks rapidly grow both in numbers and variance. This thesis is concerned with evaluation of delay strategies in anonymizing mix-networks nodes. The evaluation process includes generation of traffic data, applying different delay strategies such as uniform, exponential, batch delays, and comparison of anonymity and quality of service metrics for these strategies. Metrics of anonymity were obtained from packet delay entropies and validated with simulations. Additional simulations established clear dependency between these metrics and ability of packet mixing to mitigate attacks based on pattern recognition of packet arrival instances. |
