云计算IaaS层和SaaS层的可靠性研究

发布时间：2018-06-02 01:35

  本文选题：云计算 + 可靠性　； 参考：《内蒙古农业大学》2016年硕士论文

【摘要】：云计算为信息服务领域提供了一种有效的商业模式,它将存储和计算简化为易用资源,以按需付费的方式通过网络提供给用户。云计算具备灵活、方便、经济等多个特征,然而这些特征也让云计算与传统计算模式相比更加复杂,且更易出现故障。云计算的可靠性与云服务的性能和质量直接相关,对云计算行业的长久发展有着重要意义。本文分别对云计算IaaS层和SaaS层的故障类型和可靠性测试方法做出了主要研究,提出了IaaS层虚拟网络快照方法,并基于故障树分析方法构建了SaaS层故障树模型。具体工作如下：(1)对国内外云计算可靠性研究现状以及云计算目前存在的问题进行了概括和总结,并针对存在的各个问题类型提出了相应的解决思路。(2)分别以物理层和虚拟化层两方面对IaaS层可靠性进行研究,总结了物理层中硬件设施的故障场景、故障检测及缓解方法。同时针对虚拟化层提出了IaaS层虚拟网络快照方法,从快照图像动态迁移和快照图像相似性排查两方面提高了快照性能,有效减少了快照导致的系统停机时间,降低了快照传输时的流量消耗。(3)搭建Jelastic云平台,通过对CPU、RAM、HDD、Network的实时监控,分析总结SaaS层中各个故障类型和故障间的逻辑关系,构建SaaS层故障树模型,包含10项主要故障事件和84个基本故障事件,并通过实验求得各个主要事件的发生概率及结构重要度,为SaaS层的故障处理和改进提供参考。
[Abstract]:Cloud computing provides an effective business model for information services. It simplifies storage and computing to easy-to-use resources and is provided to users on demand through the network. Cloud computing has many characteristics, such as flexibility, convenience, economy and so on. However, these features also make cloud computing more complex and more prone to failure compared with traditional computing mode. The reliability of cloud computing is directly related to the performance and quality of cloud services, which is of great significance to the long-term development of cloud computing industry. In this paper, the fault types and reliability testing methods of cloud computing IaaS layer and SaaS layer are studied, and the IaaS layer virtual network snapshot method is proposed. Based on the fault tree analysis method, the SaaS layer fault tree model is constructed. The specific work is as follows: (1) the research status of cloud computing reliability at home and abroad and the problems existing in cloud computing are summarized and summarized. According to the existing problems, the corresponding solutions are put forward. (2) the reliability of IaaS layer is studied in physical layer and virtualization layer, and the fault scenarios, fault detection and mitigation methods of hardware facilities in physical layer are summarized. At the same time, the virtual network snapshot method of IaaS layer is proposed for virtualization layer, which improves the snapshot performance from two aspects of snapshot image dynamic migration and snapshot image similarity checking, and effectively reduces the system downtime caused by snapshot. The Jelastic cloud platform is built by reducing the traffic consumption in snapshot transmission. Through the real-time monitoring of SaaS Ram HDD Network, the logical relationship between each fault type and fault in SaaS layer is analyzed and summarized, and the fault tree model of SaaS layer is constructed. It consists of 10 main fault events and 84 basic fault events. The probability of occurrence and the structural importance of each major event are obtained through experiments, which provides a reference for the fault treatment and improvement of SaaS layer.
【学位授予单位】：内蒙古农业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.09
，

本文编号：1966710