Abstract: Application server cluster is the most important runtime environment in platform as a service (PaaS). Based on the dynamic variable load and shared heterogeneous environment in PaaS, this paper brings forward a self-adaptive approach for cluster adjusting in order to provide resource on-demand corresponding to the cluster load. The approach constructs a performance diagnosis model, then provides self-adaptive mechanism for resource provision and load balancing. The experiment shows that the approach can implement resource provision on-demand through adjusting the size of logic resource pool, the number of cluster nodes and self-adaptive load balancing.

Key words: cloud computing, self-adaptive cluster, load balancing, performance diagnosis

摘要： 应用服务器集群是平台即服务(platform as a service, PaaS)模式的主要运行环境。针对云环境下动态变化的用户负载和共享、异构的集群环境, 提出一种自适应集群调整方法, 根据集群负载状况实现资源按需供给。该方法建立了PaaS平台的性能分析模型, 并据此提出自适应的资源供给机制和负载均衡机制。实验结果表明, 通过调节集群节点逻辑资源池的大小和集群节点数量, 配合自适应负载均衡方法, 达到了资源按需供给的目的。

关键词: 云计算, 自适应集群, 负载均衡, 性能诊断