User Persistence in 5G/6G mmWave/Sub-THz Systems With Blockage: Does It Pay Off?

5G New Radio (NR) millimeter wave (mmWave) and future sub-terahertz (sub-THz) 6G systems are designed for rate-greedy multimedia applications. Such applications often utilize rate adaptation allowing to vary the rate according to the network conditions. Unreliable nature of these bands subject to blockage may induce persistent user behavior resulting in repeated attempts (retrials) to continue the service. These effects may lead to the loss of sessions during the service and service provisioning at reduced rate. In this paper, by using the stochastic geometry and queuing theory, we characterize performance of adaptive rate-greedy applications at mmWave/sub-THz base stations. We consider user- and system-centric performance including new and ongoing session drop probabilities, resource utilization, and the amount of wasted resources. Our results show that the user persistence improves the new and ongoing session drop probabilities: making two retrials brings both down by <inline-formula> <tex-math notation="LaTeX">$20-70$ </tex-math></inline-formula>% while for fully persistent behavior the gain is 300%. The impact is more pronounced in underloaded system conditions. User persistence also increases the usage of system resources by <inline-formula> <tex-math notation="LaTeX">$20-40$ </tex-math></inline-formula>% and decreases the percentage of wasted resources by two/three times. However, these gains come at the expense of decreased successful session completion probability without retrials negatively affecting the overall user experience.