基于融合特征空间的云数据离散调度逼近算法

Cloud Data Scheduling Discrete Approximation Algorithm Based on Fusion Feature Space Frame

基于融合基础架构的新一代数据中心,提出一种基于融合特征空间构架的云数据离散调度逼近算法,在对系统节点进行数据采集任务分配的基础上,研究基于云计算的IAAS层基础设施自动化快速供应的技术方案,对云数据离散分配任务进行评价和估计,描述了云数据采集和特征空间融合架构方法,构建云数据采集和特征空间融合机制,进行融合特征空间构架下的数据离散调度逼近算法改进设计。研究结果表明,该方法实现云数据离散调度的压缩性能很稳定,有效节约了任务执行时间,云数据离散调度逼近的收敛步数较少,收敛性好,展示了较强的实用性和优越性,为云计算的顶层设计PAAS与SAAS提供技术平台支持。

The traditional method is using the cloud data scheduling mechanism characteristics of stochastic approximation,starting from the single node performance, and it cannot achieve the discrete scheduling of cloud data. A cloud data fusion feature space frame scheduling algorithm is proposed based on the discrete approximation, based on data acquisition task allocation of system node, evaluate and estimate of the cloud data of discrete distribution of tasks, describes the structure method of cloud data acquisition and feature space fusion, build the cloud data acquisition and feature space fusion mechanism, fusion the feature space under the framework of data discretization scheduling approximation algorithm improved design. The research results show that the compression performance is very stable implementation of cloud data discrete scheduling the law, it can effectively save the task execution time, cloud data convergence step discrete scheduling approximation of the number, it has good convergence, it shows the strong practicality and superiority. It provide stechnical support platform of PAAS and SAAS as the top-level design of cloud computing