何元庆, 陈旻, 乐松山, 黄丙湖, 宋杰, 马载阳, 王进. 一种面向陆地碳循环模型服务的去中心化计算方法[J]. 武汉大学学报 ( 信息科学版), 2020, 45(9): 1395-1402. DOI: 10.13203/j.whugis20200203
引用本文: 何元庆, 陈旻, 乐松山, 黄丙湖, 宋杰, 马载阳, 王进. 一种面向陆地碳循环模型服务的去中心化计算方法[J]. 武汉大学学报 ( 信息科学版), 2020, 45(9): 1395-1402. DOI: 10.13203/j.whugis20200203
HE Yuanqing, CHEN Min, YUE Songshan, HUANG Binghu, SONG Jie, MA Zaiyang, WANG Jin. Service-Oriented Decentralized Computing Method for Terrestrial Carbon Cycle Model[J]. Geomatics and Information Science of Wuhan University, 2020, 45(9): 1395-1402. DOI: 10.13203/j.whugis20200203
Citation: HE Yuanqing, CHEN Min, YUE Songshan, HUANG Binghu, SONG Jie, MA Zaiyang, WANG Jin. Service-Oriented Decentralized Computing Method for Terrestrial Carbon Cycle Model[J]. Geomatics and Information Science of Wuhan University, 2020, 45(9): 1395-1402. DOI: 10.13203/j.whugis20200203

一种面向陆地碳循环模型服务的去中心化计算方法

Service-Oriented Decentralized Computing Method for Terrestrial Carbon Cycle Model

  • 摘要: 在开放式Web环境中,模型服务化共享促进了陆地碳循环模型的发展,然而海量的计算数据导致模型计算缓慢,现有模型服务化方案难以满足其计算需求。为了解决这一问题,提出了一种面向陆地碳循环模型服务的去中心化计算方法。该方法主要思想是基于现有模型服务化方案,通过增加计算节点的方法以缓解陆地碳循环模型计算压力,主要包括陆地碳循环模型服务准备和陆地碳循环模型分布式运行两个步骤。基于此设计了任务分割和派发策略将复杂计算任务离散化并派发至计算节点进行运算。将设计的方法应用到3个典型陆地碳循环模型的计算中,验证了该方法的可行性,提高了陆地碳循环模型的计算速度。

     

    Abstract: Geographic modeling and simulation is an important method for solving environmental issues and providing decision-making. Recently, research related to geographic modeling and simulation in an open web environment is mainly focused on model service sharing, while the calculation speed of model has been less studied. The terrestrial carbon cycle model plays a key role in the simulation of carbon cycle, service-oriented sharing has promoted it's development, but the time-consuming calculation speed limits their widespread applications. The large amount of calculation of terrestrial carbon cycle model results in a large amount of time to obtain the output. Currently, the model sharing strategies have not employed effective ways to improve the calculation performance of terrestrial carbon cycle models. A service-oriented decentralized computing method is proposed to speed up the calculation of these models. The main idea of this method is adding computing nodes to alleviate the pressure of the calculation of the terrestrial carbon cycle model based on the existing model sharing strategies. First, the terrestrial carbon cycle models are published as model services in the open web environment. Then, a task segmentation and dispatch strategy is designed to break the complex computing tasks into small steps and dispatch them to different computing nodes for distributed calculation. The study case of three typical terrestrial carbon cycle models show the feasibility and practicability of the proposed method and it can be used to improve the calculation speed of the terrestrial carbon cycle model.

     

/

返回文章
返回