Efficient Proactive Resource Allocation for Multi-stage Cloud-Native Microservices

Microservices deployment in the cloud often faces a prevalent challenge: how to maximize resource utilization while maintaining high quality-of-service (QoS). Existing automatic scaling tools frequently exhibit limited adaptability, particularly when handling frequent request load fluctuations, which exacerbates the challenge. To address this issue, we introduce a proactive runtime deployment optimization method for multi-stage microservices, aiming to ensure both resource efficiency and QoS. Our proposed method encompasses four interrelated modules–forecasting, constraint planning, judgment selection, and execution–which collaboratively work towards optimizing runtime resource allocation, generating viable deployment plans, and identifying cost-efficient solutions without compromising QoS. Through a set of experiments, we demonstrate that the proposed proactive deployment optimization method can potentially reduce computational resource usage by 35% while maintaining the desired quality of service.