Revisiting Controller-Based Admission Control for Virtual Machines

Edge computing has revolutionized computational paradigms by bringing computing resources closer to the end-user, significantly reducing latency. This shift is crucial for latency-sensitive applications and services, which cannot tolerate the delays caused by transmitting workloads to remote cloud data centers. This study delves into the influence of server computation on Quality of Service (QoS), with a particular emphasis on how overbooking CPU resources affects latency. It emphasizes the finite conservative capacity of servers, exceeding this capacity can impair the performance of hosted applications. This research revisits prior studies on control theory for admission control and validates the practical application of the Proportional-Integral-Derivative (PID) controller in managing Virtual Machine (VM) admissions in a virtualized server environment. Furthermore, it also presents a comparative analysis of PID, Proportional-Derivative (PD), Proportional-Integral (PI), and on/off controllers, offering insights into their effectiveness in optimizing server performance.