基于灰色模型的云资源动态伸缩功能研究

Research on Cloud Resources Auto-scaling Based on Grey Model

随着云计算的快速发展,大量应用开始向云上迁移,云平台资源分配受到关注。云服务负载是动态变化的,为保证服务质量不受负载量变化影响,最大化利用云资源,如何动态扩充、缩紧资源成为需要考虑的重要问题。设计并实现了一个Docker Swarm弹性云动态伸缩模块,通过增加和减少云服务副本数量改变服务资源的分配。弹性云的动态伸缩模块使用了响应式伸缩模型与预测式伸缩模型,其中预测式伸缩模型基于灰色预测模型。实验证明基于灰色预测的预测式伸缩模型具有较高的预测准确率。

With the development of cloud-computing,more and more application start to move on cloud,and the resource allocation of cloud platform has been much accounted.The workload of cloud-service is changing dynamically.It is the main concern of how to maximize cloud-resource and realize auto-scaling to ensure quality service.A Docker Swarm elastic-cloud platform has been designed and achieved,which uses replica to change resource allocation of platform.The auto-scaling model in elastic cloud uses both reactive scaling model and prediction scaling model to ensure auto-scaling.The prediction scaling model is based on Grey-forecasting model.The experiment proves the prediction scaling model which based on Grey-forecasting model has high prediction accuracy.