基于动态环境的移动云计算切分方法的研究

发布时间：2018-02-21 12:23

本文关键词： 移动云计算单帧数据流计动态环境计算切分数据流应用性能解析　出处：《湖北工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：在中国这个高速发展的社会中,互联网发展的也是相当的迅速,云服务就是互联网发展的产物,其中移动云计算是云服务系统中的移动应用产生的一种新的云计算方式。电池容量、网络通信能力和计算存储资源等是移动终端设备本身具有局限性,移动云计算有其局限性,但也有其有点:在无线网络下,可以使移动终端设备拓展移动应用,而且还可以使用云计算的资源,按照需求获取存储资源及所需的计算。移动云计算为了能够提高移动应用的服务质量(Quality of Service,Qos)会把全部或部分需要计算的工作移动到云端。在当前的移动云计算方式还较少,要探究适合移动客户端随处境变化和能用资源网络宽带的状况下去提升功能。对于上面所讲的问题,在本篇文章中会讲述关于移动云服务应用的计算切分方法在动态环境下的移动云计算的基础上的研究,在动态环境的影响下去探究计算工作单元在本地计算或者移动到云端计算情况下的动态决策移动应用中的计算。主要展开下面所讲任务:在与移动云计算系统框架相结合的情况下,最先建立一种有状态的移动数据流应用的模型。对于多帧数据流计算切分决策、任务调度和执行效率、动态优化及单帧数据流计算切分方案等问题的定义解析是探究移动云计算的问题和方向定义所包括的几个方面的内容。单帧数据流计算切分方法是在以上基础探究的并在知道计算切分方案的情况下提出工作排序及工作任务数据流模型,而且解析了它的工作调度办法,计算数据帧经过计算切分后的执行时间是依据调度办法计算的,并且经过遗传的计算方法来执行最短的时间下作为适应度函数,进而计算出最佳的计算切分决策办法。本篇文章进行多帧数据的操纵是在单帧数据流计算切分办法的基础上进行的,去探究多帧数据优化调整及计算切分决策,考虑到网络带宽在未来短时间内的变化情况,要实现多帧数据计算切分方案优化决策是经过计算节点的计算切分方案的优化调整办法来实现的。对于上面探究的工作内容,本篇文章考证阐述了有状态的数据流应用中单帧数据计算切分的有效性并分析了多帧数据在优化调整后的方法的性能。在本篇论文的最后,考证论文办法的有效性是经过一个移动云计算平台原型系统“移动云服务人脸识别”来验证的。在伴随着移动设备终端性能的提高级4G时代的的到来,本篇文章探究的移动云计算的计算切分方法能够更广泛地应用到像语音识别处理系统及条码二维码识别等图像识别处理系统等有状态的数据流处理程序中。
[Abstract]:In this rapidly developing society in China, the Internet is also developing quite rapidly. Cloud services are the product of the development of the Internet. Mobile cloud computing is a new way of cloud computing generated by mobile applications in cloud service systems. Battery capacity, network communication capacity and computing storage resources are the limitations of mobile terminal devices, and mobile cloud computing has its limitations. But it's also a little bit: in wireless networks, mobile devices can expand their mobile applications, and they can use cloud computing resources. Mobile cloud computing moves all or part of the work needed to compute to the cloud in order to improve the quality of service of mobile applications. To explore the mobile client to adapt to the changing situation and the ability to use resource network broadband to upgrade the function. In this article, we will talk about the research of computing segmentation of mobile cloud service application based on mobile cloud computing in dynamic environment. Under the influence of the dynamic environment to explore the computing work unit in local computing or moving to cloud computing dynamic decision-making mobile application computing. The main tasks described below: with the mobile cloud computing system box. When the frame is combined, A model of stateful mobile data stream application is first established. For multi-frame data stream computing segmentation decision, task scheduling and execution efficiency, The definition analysis of dynamic optimization and single-frame data stream computing segmentation scheme is to explore the problems of mobile cloud computing and several aspects of the definition of direction. The single-frame data stream computing segmentation method is based on the above. The work order and work task data flow model are proposed under the condition of knowing the calculation of the segmentation scheme. Moreover, its scheduling method is analyzed. The execution time of the calculated data frame is calculated according to the scheduling method, and the genetic calculation method is used to perform the fitness function in the shortest time. In this paper, the manipulation of multi-frame data is carried out on the basis of single-frame data stream calculation and segmentation, to explore the optimization adjustment of multi-frame data and the calculation of segmentation decision. Considering the change of network bandwidth in a short time in the future, the optimization decision of multi-frame data computing segmentation scheme is realized through the optimization adjustment method of computing node's computing segmentation scheme. This paper discusses the validity of single-frame data segmentation in the application of stateful data flow and analyzes the performance of the method after optimization and adjustment of multi-frame data. The validity of this method is verified by a prototype system of mobile cloud computing platform "mobile cloud service face recognition". The computational segmentation method of mobile cloud computing explored in this paper can be more widely used in stateful data stream processing programs such as speech recognition processing systems and image recognition processing systems such as barcode two-dimensional code recognition.
【学位授予单位】：湖北工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09

【相似文献】