基于预先分类的分布式水下网络空间多维数据并行调度方法

A pre-classification based parallel scheduling method for multidimensional data in distributed underwater networks

水下网络数据调度方法存在网络空间多维数据调度策略异常、交互资源槽分布调度存在资源均衡分布异常的状况,导致数据交互调度能力下降,数据间交互延迟增大,多维数据类别化调度准确率受到影响。针对问题提出基于预先分类的分布式水下网络空间多维数据并行调度方法,首先,对调度数据类别进行资源槽的类别优化处理,通过引入资源槽与分类调度算法,理顺资源槽交互类别数据集;接着对网络空间多维数据进行调度逻辑的计算,根据网络数据传输特点,引入多维数据分布式云并行调度算法,对网络空间中的并行数据流进行优化,实现多维数据的并行调度;最后,通过设计1000~2000组的实验数据,对提出方法的可行性进行证明,证明方法具有可行性高、数据并行调度效率高、稳定好的特点。

Underwater network data scheduling method has abnormal multi-dimensional data scheduling strategy in network space and abnormal balanced distribution of resources in interactive resource slot distribution scheduling,which results in the decline of data interactive scheduling ability,the increase of interaction delay between data,and the impact of multi-dimensional data classification scheduling accuracy.Aiming at the problem,a distributed parallel scheduling method for multi-dimensional data in underwater network space based on pre-classification is proposed.Firstly,resource slots are optimized for scheduling data categories.By introducing resource slots and classification scheduling algorithm,the interactive class data sets of resource slots are straightened out.Secondly,multi-dimensional data in network space are scheduled.Logical calculation,according to the characteristics of network data transmission,introduces the multi-dimensional data distributed cloud parallel scheduling algorithm,optimizes the parallel data flow in network space,and realizes the parallel scheduling of multi-dimensional data.Finally,through designing 1000~2000 groups of experimental data,the feasibility of the proposed method is proved and the method is proved.It has the characteristics of high feasibility,high data parallel scheduling efficiency and good stability.