Study On Joint Computing Model Of RTM Based On CESM

The change of surface runoff is the main factor to change the ecological environment,especially the regional climate,and the RTM module of CESM is an effective way to study the surface runoff,but it needs great calculation.The rarity of supercomputer resources is in contradiction with the demand of the extensive and multi diversity research.To solve this problem,we need to use computer cluster to get computing power in research.Many scientific studies are based on high performance computing clusters.However,many technologies have been developed in the distributed computing of clusters,such as grid computing,meta computing,etc.Grid computing can link the idle computing devices in different locations in the network to form a computing cluster.However,how to make the cluster perform its best performance and shorten the computation time is a difficult problem.This needs to be found in a heterogeneous or isomorphic network cluster,which can be split,distributed,recycled,integrated,and the load balancing algorithm of each resource node is also good.For grid cluster computing,task scheduling and load balancing have always been the focus of research.The focus of task scheduling and load balancing is to understand the self attributes and state information of each resource node in the network,so as to avoid the unbalance of the load which is too centralized or the node is too idle.In this way,the optimal scheduling strategy can be realized.Therefore,the following contents are studied in this paper.In this paper,the grid task scheduling technology and load balancing technology are introduced in this paper,and the important role of task scheduling and load balancing in grid and cluster networks is explained in detail.In view of the characteristics of RTM numerical calculation,combined with many aspects of ant colony algorithm,the ant colony algorithm is used to expand its characteristics of ant foraging.The ant colony algorithm is introduced in the grid system,and the ant colony algorithm is improved properly so that it can meet the requirements of the RTM numerical calculation and allocate the tasks more effectively to the matching nodes and shorten the time of the grid task.In order to make cluster network management more flexible and improve resource utilization,avoid network congestion and enhance network transmission performance.In this paper,the SDN is used to deploy the cluster network,and the ant colony algorithm is introduced.According to the aim of this paper,the parameters of the ant colony algorithm are adjusted properly,so that the optimal path can be planned in the SDN network,and the node resource equipment can be efficiently used to improve the performance of the cluster computing.