基于负载均衡的农业信息服务云平台的设计

发布时间：2018-07-28 11:00

【摘要】：在中国农业信息化的发展过程中,大量的农业知识、数据、资源、技术和市场需求等信息无法在农村地区广泛迅速地传播,基层农业人员无法获得有价值的相关信息,即存在“最后一公里”现象。针对这一问题,本文提出了一种基于OpenStack的农业信息服务云平台架构。该云平台面向农业科研机构、农业企事业单位及农户个人提供农业信息发布、信息查询和高性能的计算、存储等软硬件资源共享服务。本文的主要工作如下:首先,提出了基于Openstack的农业信息服务云平台架构。在调研国内外农业信息化的研究现状的基础上,结合我国农业信息化需求的特点,提出了农业信息服务云平台的设计架构。该平台采用系统层次结构,分为基础层、应用服务层、管理层和接入层。层与层之间使用标准的接口设计,实现单点登录,统一管理。在已有的设施基础上,部署Openstack,搭建综合服务云平台,从而实现农业相关资源的整合、农业信息的发布和检索、农业生产管理和交互服务。然后,对于农业信息服务云平台的资源调度策略进行改进。针对农业信息服务的复杂性、多样性、实时性的特征,对OpenStack的原有调度策略进行改进,提出了一种基于虚拟机迁移的动态资源调度策略,使农业信息服务云平台能更好地适应农业信息负载的动态变化。最后,对于改进的调度算法性能进行实验仿真验证。利用Cloudsim仿真软件,对比改进后的调度算法与常用的调度算法。实验结果表明,改进的调度算法能较好地保证负载均衡,同时节能效果良好,在不违反服务等级协议(SLA)的前提下,数据中心的整体能耗和响应时间都明显降低。
[Abstract]:In the course of the development of China's agricultural informatization, a large amount of agricultural knowledge, data, resources, technology and market demand can not be widely and rapidly disseminated in rural areas, and grass-roots agricultural personnel cannot obtain valuable relevant information. That is, the existence of the "last kilometer" phenomenon. To solve this problem, this paper proposes a cloud platform architecture of agricultural information service based on OpenStack. The cloud platform for agricultural research institutions, agricultural enterprises and institutions and individual farmers to provide agricultural information release, information query and high-performance computing, storage and other hardware and software resources sharing services. The main work of this paper is as follows: firstly, the architecture of agricultural information service cloud platform based on Openstack is proposed. On the basis of investigating the present situation of agricultural informatization at home and abroad and combining the characteristics of agricultural informatization demand in our country, the design framework of agricultural information service cloud platform is put forward. The platform adopts system hierarchy, which is divided into basic layer, application service layer, management layer and access layer. Layer and layer between the use of standard interface design, single sign-on, unified management. On the basis of the existing facilities, Openstackand build an integrated service cloud platform to realize the integration of agri-related resources, the release and retrieval of agricultural information, agricultural production management and interactive services. Then, the resource scheduling strategy of agricultural information service cloud platform is improved. In view of the complexity, diversity and real-time characteristics of agricultural information service, the original scheduling strategy of OpenStack is improved, and a dynamic resource scheduling strategy based on virtual machine migration is proposed. The cloud platform of agricultural information service can better adapt to the dynamic change of agricultural information load. Finally, the performance of the improved scheduling algorithm is verified by simulation. By using Cloudsim simulation software, the improved scheduling algorithm and the common scheduling algorithm are compared. The experimental results show that the improved scheduling algorithm can ensure load balance and save energy well. Without violating the service level protocol (SLA), the overall energy consumption and response time of the data center are obviously reduced.
【学位授予单位】：河南师范大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP311.52;S126

【参考文献】