In this paper, a mobile virtual reality (VR) video streaming system is considered and assessed. The system comprises of a VR server, a VR center offering mobile edge computing along with a buffer, and a mobile head-mounted display (HMD). Markov-modulated processes are used to model the arrival and departure traffic at the VR center and the time-varying link quality of the wireless channel to the mobile HMD. As the Markovian arrival process at the VR center is independent of the link quality variations of the wireless channel, the VR center controls the delivery of VR video packets to the mobile HMD subject to the link quality of the wireless channel. The Markov models are used to derive expressions of queueing theoretical performance measures, i.e., throughput, queueing delay, and packet loss probability. Numerical results are provided to investigate performance trade-offs for real data traces. © 2024 IEEE.