基于云存储的视频信息分布式优化处理系统的研究与设计
发布时间：2019-06-05 12:21
【摘要】：随着科学技术的进步,视频处理系统虽然得到一定发展,但面对需要满足大量访问、快速响应等高质量视频服务时,传统方案在体系结构和负载均衡等方面还不够成熟,已经不能再满足当今的需求。本文所讨论的基于云存储的视频信息分布式优化处理系统,其中云存储作为一个新兴的研究和应用领域,其具有快速部署,低成本,灵活调整规模等优势,但云存储同样也受到了一定的限制,原因在于我们虽然拥有一系列对负载均衡进行衡量的算法,但是由于不能提前对负载进行预算度量,这就使负载均衡失去了基础,限制了整个系统性能。基于小波神经网络的负载均衡具有可预测性和自学习性,使负载均衡达到合理应用性。 本文就如何构建云存储环境、如何优化视频信息处理技术和如何运用小波神经网络来处理负载均衡策略这三个方面,给出了基于云存储的视频信息分布式优化处理系统的设计方案,在某种程度上解决传统视频信息处理系统技术上的不足,大大简化其应用环节,实现视频信息资源充分共享,提高其利用效率。本课题的主要研究工作如下： (1)云存储构架的研究与设计。基于云存储概念及特点,设计了云存储四层存储服务器模型,从底层到上层依次是：云存储层,数据管理层,应用接口层(也叫数据服务层)以及用户访问层。本文提供的设计方案为：利用普通PC机群搭建云存储中的底层-云存储层,采用多种功能模块分块管理进行数据管理层的设计,在应用接口层针对相应功能开发一些实际接口,方便与访问层用户操作的交互。 (2)视频信息分布式优化处理。基于云存储环境,系统分别从视频信息传输、调度、存储等方面进行优化设计。对接收到的视频信息进行重组及H.264解码,采用TCP与RTP相结合的方式进行传输。在调度方面,选择一种新调度算法-最强能力优先调度算法,存储策略则是采用基于时间序列的视频文件热度进行有效存储。 (3)负载均衡的研究与设计。针对传统算法的局限性,文章提出了一种基于小波神经网络预测模型的改进算法,并在MATLAB环境下进行仿真实验,证明优越性。并以此为基础,设计了系统负载均衡策略。
[Abstract]:With the progress of science and technology, although video processing system has been developed to a certain extent, the traditional scheme is not mature enough in architecture and load balancing when it needs to meet a large number of access, fast response and other high-quality video services. It can no longer meet the needs of today. The distributed optimal processing system of video information based on cloud storage discussed in this paper, in which cloud storage, as a new research and application field, has the advantages of rapid deployment, low cost, flexible adjustment of scale and so on. But cloud storage is also limited because although we have a series of algorithms to measure load balancing, because we can not measure the load ahead of time, load balancing has lost its foundation. The performance of the whole system is limited. Load balancing based on wavelet neural network has predictability and self-study habit, which makes load balancing reach reasonable application. This paper focuses on how to construct cloud storage environment, how to optimize video information processing technology and how to use wavelet neural network to deal with load balancing strategy. The design scheme of video information distributed optimization processing system based on cloud storage is given, which solves the technical shortcomings of traditional video information processing system to some extent, greatly simplifies its application links, and realizes the full sharing of video information resources. Improve its utilization efficiency. The main research work of this paper is as follows: (1) the research and design of cloud storage architecture. Based on the concept and characteristics of cloud storage, a four-layer storage server model of cloud storage is designed. From the bottom layer to the upper layer, the cloud storage layer, the data management layer, the application interface layer (also known as the data service layer) and the user access layer are designed. The design scheme provided in this paper is as follows: using ordinary PC cluster to build the bottom layer of cloud storage, using a variety of functional modules to manage the data management, and developing some practical interfaces for the corresponding functions in the application interface layer. Facilitate interaction with access layer user operations. (2) distributed optimal processing of video information. Based on cloud storage environment, the system is optimized from the aspects of video information transmission, scheduling, storage and so on. The received video information is reorganized and H.264 decoded, and TCP and RTP are combined to transmit the video information. In the aspect of scheduling, a new scheduling algorithm, the strongest capability priority scheduling algorithm, is selected, and the storage strategy is to store the video file heat based on time series effectively. (3) Research and design of load balancing. Aiming at the limitation of traditional algorithm, this paper proposes an improved algorithm based on wavelet neural network prediction model, and carries on the simulation experiment in MATLAB environment to prove the superiority. On this basis, the load balancing strategy of the system is designed.
【学位授予单位】：广东工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP391.41;TP333
本文编号：2493525
[Abstract]:With the progress of science and technology, although video processing system has been developed to a certain extent, the traditional scheme is not mature enough in architecture and load balancing when it needs to meet a large number of access, fast response and other high-quality video services. It can no longer meet the needs of today. The distributed optimal processing system of video information based on cloud storage discussed in this paper, in which cloud storage, as a new research and application field, has the advantages of rapid deployment, low cost, flexible adjustment of scale and so on. But cloud storage is also limited because although we have a series of algorithms to measure load balancing, because we can not measure the load ahead of time, load balancing has lost its foundation. The performance of the whole system is limited. Load balancing based on wavelet neural network has predictability and self-study habit, which makes load balancing reach reasonable application. This paper focuses on how to construct cloud storage environment, how to optimize video information processing technology and how to use wavelet neural network to deal with load balancing strategy. The design scheme of video information distributed optimization processing system based on cloud storage is given, which solves the technical shortcomings of traditional video information processing system to some extent, greatly simplifies its application links, and realizes the full sharing of video information resources. Improve its utilization efficiency. The main research work of this paper is as follows: (1) the research and design of cloud storage architecture. Based on the concept and characteristics of cloud storage, a four-layer storage server model of cloud storage is designed. From the bottom layer to the upper layer, the cloud storage layer, the data management layer, the application interface layer (also known as the data service layer) and the user access layer are designed. The design scheme provided in this paper is as follows: using ordinary PC cluster to build the bottom layer of cloud storage, using a variety of functional modules to manage the data management, and developing some practical interfaces for the corresponding functions in the application interface layer. Facilitate interaction with access layer user operations. (2) distributed optimal processing of video information. Based on cloud storage environment, the system is optimized from the aspects of video information transmission, scheduling, storage and so on. The received video information is reorganized and H.264 decoded, and TCP and RTP are combined to transmit the video information. In the aspect of scheduling, a new scheduling algorithm, the strongest capability priority scheduling algorithm, is selected, and the storage strategy is to store the video file heat based on time series effectively. (3) Research and design of load balancing. Aiming at the limitation of traditional algorithm, this paper proposes an improved algorithm based on wavelet neural network prediction model, and carries on the simulation experiment in MATLAB environment to prove the superiority. On this basis, the load balancing strategy of the system is designed.
【学位授予单位】：广东工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP391.41;TP333
【引证文献】
相关硕士学位论文 前1条
1 陈聪;基于云存储的视频监控平台[D];华南理工大学;2012年
，本文编号：2493525
本文链接：https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2493525.html
