基于云计算的视频转码自适应方法研究

发布时间：2018-06-15 17:11

本文选题：自适应视频流 + 移动云计算　；参考：《南京邮电大学》2017年硕士论文

【摘要】：近年来,云计算领域新技术层出不穷且呈现不断融合的趋势,云媒体技术日益成熟,提供高质量多样化的多媒体解决方案成为了可能。如何满足用户对在存储、计算能力有限以及电量有限的手持设备上享受高质量和多样化云多媒体服务的需求仍然是一项有趣和具有挑战性的研究。怎样合理并且高效地利用云资源及移动终端资源,实现低成本、高能效并满足服务质量需求的资源供应和服务提供,是云视频服务系统中不可避免的且具有实际意义的热点问题。本文结合云计算技术和传统的自适应视频流技术,对基于云计算的自适应视频流传输框架、移动终端高能效的分辨率码率调度策略、成本优化和服务质量(Quality of Service,QoS)保证的动态资源调度三个方面进行了研究,主要工作如下:(1)研究基于云计算的自适应视频流传输系统的架构设计问题。传统的DASH(Dynamic Adaptive Streaming over HTTP)系统实现自适应视频流所需要的计算以及缓存工作都必须在终端上进行,终端设备的能耗巨大。因此本文借助云计算技术强大的计算能力和海量存储能力,把部分计算迁移至云计算中心,研究基于云计算的自适应视频流传输系统的设计。系统实现了把移动终端的信息实时传输到云端,云端根据移动终端的网络能力和实时能耗进行策略评估,并对原始视频进行实时转码后再传送到移动终端。(2)在系统框架的基础上,研究基于终端能耗感知的分辨率码率自适应算法。参考了现有研究成果,首先对智能手机主要耗电部件进行逐一详细的分析,并进一步分析了视频流式处理过程中各个环节智能手机的能耗情况;其次建立了智能终端设备耗电因素和视频编解码参数之间的数学模型;最后提出了改进的基于终端能耗感知的分辨率码率自适应算法以求解数学模型。实验结果表明,相比DASH自适应视频流策略,本文算法能够依据终端设备能耗变化来自适应的调节视频分辨率和码率,有效降低了终端设备的能耗。(3)在系统框架基础上,研究面向云端视频转码的基于成本的资源自适应配置算法。实现自适应视频流技术必然离不开云环境下的实时转码,本文针对海量视频转码需求给视频服务供应商带来的巨大资本压力问题,首先建立QoS模型;其次,将IaaS云平台的虚拟机动态供应问题建模成实时转码过程中需要激活的虚拟机数量的最小化问题;最后通过本文提出的成本优化和服务质量保证的自适应资源调度策略对虚拟机的数量进行动态调整。实验结果表明,本文所提出的算法相比传统的资源调度方法不仅降低了视频服务供应商的成本和服务器过载概率,还提高了资源平均利用率和服务质量。
[Abstract]:In recent years, new technologies in the field of cloud computing emerge in endlessly and show a trend of continuous integration. Cloud media technology is becoming more and more mature, and it is possible to provide high-quality and diversified multimedia solutions. How to meet users' demand for high-quality and diversified cloud multimedia services on handheld devices with limited storage, computing power and limited power remains an interesting and challenging study. How to make rational and efficient use of cloud resources and mobile terminal resources to achieve low cost, high energy efficiency and meet the quality of service demand for resource supply and service delivery, It is an inevitable and practical hot issue in cloud video service system. Combined with cloud computing technology and traditional adaptive video stream technology, this paper analyzes the adaptive video stream transmission framework based on cloud computing and the high resolution rate scheduling strategy for mobile terminals. Cost optimization and dynamic resource scheduling guaranteed by quality of Service (QoS) are studied. The main work is as follows: 1) the architecture design of adaptive video streaming system based on cloud computing is studied. The traditional DASHN dynamic Adaptive streaming over (DASH) system has to perform the computing and caching work on the terminal for adaptive video stream, and the terminal equipment has a huge energy consumption. Therefore, with the powerful computing power and mass storage capacity of cloud computing technology, this paper migrates part of computing to cloud computing center, and studies the design of adaptive video stream transmission system based on cloud computing. The system realizes the real-time transmission of mobile terminal information to the cloud. The cloud evaluates the strategy according to the network capability and real-time energy consumption of the mobile terminal. The original video is transcoded in real time and then transmitted to the mobile terminal. On the basis of the system framework, the resolution rate adaptive algorithm based on terminal energy consumption sensing is studied. Referring to the existing research results, the main components of smart phone power consumption are analyzed one by one, and the energy consumption of each link in the process of video flow processing is further analyzed. Secondly, the mathematical model between the power consumption factor of intelligent terminal equipment and the video coding and decoding parameters is established. Finally, an improved adaptive algorithm based on terminal energy consumption perception is proposed to solve the mathematical model. The experimental results show that compared with the DASH adaptive video stream strategy, the algorithm can adaptively adjust the video resolution and bit rate according to the energy consumption change of the terminal equipment, and effectively reduce the energy consumption of the terminal equipment on the basis of the system framework. A cost-based adaptive resource allocation algorithm for cloud video transcoding is studied. The realization of adaptive video stream technology is bound to be inseparable from the real-time transcoding in the cloud environment. This paper aims at the huge capital pressure caused by the mass video transcoding requirements to the video service providers, first, establishes the QoS model; secondly, The dynamic provisioning of virtual machines in the IaaS cloud platform is modeled as the minimization of the number of virtual machines that need to be activated in the real-time transcoding process. Finally, the number of virtual machines is dynamically adjusted by the adaptive resource scheduling strategy of cost optimization and quality of service assurance proposed in this paper. Experimental results show that the proposed algorithm not only reduces the cost of video service providers and server overload probability, but also improves the average resource utilization and quality of service compared with the traditional resource scheduling method.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.8

【参考文献】