Dynamic Scalability Mechanisms for Microservices in Federated Cloud Platform

In recent years, cloud platforms tailored for the development of microservices applications have garnered significant attention. Traditional cloud computing platforms suffer from remoteness from end-users and unreliable network connectivity, making them ill-suited for meeting real-time task requirements. Moreover, the complex and dynamic nature of cloud computing environments, coupled with the collaborative nature of microservices applications, introduces a substantial risk of service failures in cloud services. Existing cloud platforms primarily rely on service load elasticity to mitigate service failures but fail to consider the dynamic nature of federated cloud clusters, resulting in service oscillations and delayed fault handling, thereby compromising the reliability of microservices applications in federated clouds. To address these challenges, we propose a dynamic microservices adjustment framework based on the MAPE control model. The framework regularly collects runtime service data and proactively identifies and adjusts abnormal services. It employs a dynamic window-based elasticity approach, enhanced by fast scaling and slow shrinking adjustment strategies, to enhance service stability in the face of fluctuating workloads. In real-world operational scenarios, experimental results demonstrate the effectiveness of the proposed approach compared to traditional methods.