Energy-delay-secrecy tradeoffs in wireless communications under channel uncertainty

In this paper, energy-efficient and secure communication over wireless fading channels is investigated when data arrivals are modeled as Markovian processes (e.g., discrete Markov and Markov fluid processes) in the presence of limitations on buffer overflow probabilities. It is assumed that the transmitter knows only the distributions of the fading coefficients in both the eavesdropper's and legitimate user's channels. Consequently, data transmissions occur at fixed rates, with secrecy outage happening when the instantaneous secrecy capacity is less than the fixed rate. In this setting, by employing the effective bandwidth and effective capacity formulations, throughput expressions are derived, and energy efficiency metrics, namely minimum energy per bit and wideband slope, are determined. Overall, the impact of source and channel variations as well as the delay quality of service (QoS) requirements on the energy efficiency of secure communications is identified.