Queueing-Based Latency Minimization in Video Streaming Systems with Adaptive Service Control

Abstract

Maintaining low latency and uninterrupted playback is a major challenge in video streaming systems operating under varying traffic loads and server failures. A bulk service queueing model is developed in which requests arrive in batches according to a Poisson process and are served under a general service-time distribution. The server dynamically switches among regular, slow, and fast service modes according to queue congestion and failure conditions, while vacation and dormant states represent periods of low demand. The supplementary variable technique is used to derive the probability generating function of the queue length and obtain key performance measures, including mean queue length, waiting time, busy period, idle period, and server utilization. A threshold-based cost analysis is employed to examine the trade-off between system performance and resource utilization. Numerical results demonstrate that adaptive service control reduces latency and improves system responsiveness under changing workloads. The proposed framework provides useful guidance for the design of reliable and latency-sensitive video streaming systems.