P2P网络多视点立体视频传输的数据块调度策略

立体视频数据量大、编码结构复杂,因此视频数据块的重要性和紧急程度对视频的播放质量有更大的影响作用。综合考虑立体视频播放质量和网络传输效率,提出一种联合优先级和随机传输的P2P数据块调度算法。依据数据块的重要性、紧急性和稀有性计算传输优先级别,同时评估邻居节点传输能力,利用多余带宽传输随机数据块。该算法能够提供优越的视频播放质量,并能够使数据块均匀分布,保证系统传输性能。仿真实验通过分析视频PSNR、启动延迟、节点吞吐量和数据到达率等指标,证明了该算法的有效性。     

一种自适应的P2P流媒体数据调度算法

针对现有算法在有限带宽情况下数据块易丢失的问题,提出了一种新颖的数据调度算法——最少优先调度算法（LFSA）。该算法易于实现,并且可以根据网络的异构性,带宽抖动以及P2P网络节点的动态性进行自动调整。实验结果表明,LFSA能够有效缩短系统播放启动延迟,而且可以在不知道伙伴节点带宽和没有任何负载平衡机制的情况下取得比传统策略更好的性能。     

一种基于节点可选度的P2P点播流媒体数据调度算法

当前的P2P(Peer-to-Peer)点播流媒体系统中数据调度算法未能充分利用每个用户节点自身的特性.在分析典型数据调度算法基础上提出一种基于节点可选度的数据调度算法(SSP算法).该算法一方面在调度下载数据块时综合考虑了邻居节点带宽能力及其所拥有的数据信息.另一方面对服务节点的请求处理过程进行了优化.SSP算法有利于提高用户节点播放视频的连续性,降低流媒体服务器的负载压力,从而改善P2P点播流媒体系统的整体服务质量.仿真结果和实际应用表明算法性能良好,适用于用户节点能力差异较大的P2P点播流媒体环境.     

P2P流媒体的数据调度算法

数据驱动型覆盖网络中的数据调度算法是影响P2P流媒体系统性能的重要因素,为了解决当前算法未能有效利用数据块和节点的特性导致流媒体服务质量差的问题,提出了一种基于数据块优先级和节点能力度的数据调度算法。该算法能够根据数据块的稀缺性、紧迫性得到块的优先级大小,根据节点的上行带宽、在线时间、相对距离得到节点能力度的大小,使优先级高的数据块和能力度大的节点优先被请求,减少了数据块的播放等待时间。在OPNET网络仿真实验表明该算法能够有效降低流媒体播放系统的启动延时和服务器的负载。     

基于贡献度的自适应P2P流媒体数据调度算法

在P2P流媒体系统中,数据调度算法是决定视频播放质量和系统性能的核心部分。针对当前P2P流媒体数据调度算法未能考虑节点带宽和服务能力的差异,从而造成对系统资源利用不充分的问题,提出一种基于贡献度的自适应(CBA)流媒体数据调度算法。算法定义节点贡献度来衡量节点的数据上传和可用带宽情况,预先向部分节点传输准备数据。并根据数据块优先级、带宽估计情况和节点贡献度等信息进行自适应调整,确定数据块请求的提供方和次序。仿真实验表明,CBA算法能充分地利用节点可用带宽,降低流媒体的启动延迟和服务器负载,改善系统的整体性能。     

基于综合因素的P2P流媒体数据调度算法

为了提高P2P流媒体数据传输的效率,提出一种基于综合因素的P2P流媒体数据调度算法,该算法综合考虑数据块的提供者数量和紧急程度定义数据块优先级,并对与提供者之间的网络带宽进行有效估计,根据优先级和带宽调度P2P数据请求的次序和方向。仿真结果表明,该调度算法在节点吞吐量和系统连续性指标以及启动延迟方面比传统的P2P流媒体调度算法有明显提高。     

上下文感知的自适应P2P流媒体数据调度算法

在网格型P2P流媒体系统中,媒体内容通常分成不同的数据块。其中,负责协调来自多个发送节点的数据块的调度算法,是影响用户感知的视频质量的重要因素。为了提高流媒体系统的整体性能,提出一种上下文感知的自适应(CAA)流媒体数据调度算法。算法根据上下文信息定义了数据块的优先级,并动态估计与邻居节点间的网络带宽,根据数据块的优先级、发送节点质量的评估和网络容量等上下文信息计算数据块请求的次序和方向。仿真结果表明,CAA调度算法具有较小的缓冲延迟,在节点吞吐量和系统负载均衡方面比传统的P2P流媒体调度算法有所提高,同时节点连续性指标也得到了改进。     

一种基于激励机制的P2P流媒体数据调度模型

针对P2P流媒体中现有算法在数据调度时不够公平的问题,提出了一种新的具有激励作用的算法模型.该算法模型可以根据节点的丢包率、贡献效率动态调节节点带宽分配以达到接收节点利益理性最大化和系统整体公平,提高了系统传输效率.仿真实验表明,该算法模型促进了节点的参与传输调度,达到了维持系统公平的目的.     

基于请求响应控制机制的数据调度策略

现有P2P流媒体系统的调度策略大多在请求端进行数据调度,在响应端则采用简单的先来先服务或最小时限优先响应的响应策略,其缺点是优先级较高的请求在响应时未必被优先响应。为此,综合考虑数据块的紧迫性、稀缺性以及请求节点的带宽,提出一种基于请求响应控制机制的数据调度策略(DSBRC)。与现有的数据调度策略相比,DSBRC提高系统上行带宽利用率,同时缩短节点启动时延和提高节点播放质量。     

基于缓存偏移量的P2P流媒体调度策略

现有P2P流媒体调度策略在确定数据块调度优先级时,未能综合考虑数据块稀缺度和紧迫度因素,导致了较差的播放连续度和启动延迟性能。提出的基于缓存偏移量的数据调度策略通过统计数据块在供求节点上的播放偏移量并对其加以处理,可有效懈决该问题,使播放连续度、启动延迟性能得到明显提高。     

一种分阶段的P2P流媒体数据调度算法

针对当前P2P流媒体中各节点数据调度时独立获取数据而对提高节点间数据协作性考虑不足,导致服务器负载较重的问题,提出一种分阶段的数据调度算法。当多个节点将请求的数据中含有相同数据块时,各节点将分两个阶段合作的获取这些数据。第1阶段按照带宽相对大小各节点以合作的方式先各自获取数据的一个子集,第2阶段节点间再充分共享上一阶段已经获得的子集数据,协作地完成该轮调度的数据传输。仿真实验表明该算法可以提高节点间数据协作性,充分利用节点带宽资源传输媒体数据,提高系统可扩展性。     

基于改进粒子群算法的P2P流媒体数据调度策略

根据P2P流媒体数据调度特点,改进了粒子群优化算法,并提出适用于离散粒子群算法（MDPSOA）的数字串编码方式。调度策略引入资源紧急度以及资源稀缺度来选择调度数据片,然后用改进离散粒子群算法来进行节点寻优,找出最优调度节点集。最后通过实验仿真算法收敛性、数据调度时间、网络利用带宽和节点负载平衡,从而验证调度策略的可行性和有效性。     

对等网络副本最佳占有率缓存管理策略

随着对等网络应用的不断深入,如何减少时间延迟,减轻集中性带宽负载,提高服务质量,已经成为研究的一个重点.提出了CORPC缓存管理方案.该方案通过使用流媒体片段的流行度来定义媒体片段副本数可占用的最佳系统缓存容量,综合考虑流媒体片段已有的副本容量、流媒体片段的热度、系统节点存储容量,使用启发式贪婪算法来实现缓存准入和缓存替换机制.该方案兼顾了不同热度的媒体片段的服务质量.模拟环境的测试结果表明,随着节点缓存空间的增加,系统服务质量得到改善.     

An efficient playout smoothing mechanism for layered streaming in P2P networks

Layered video streaming in peer-to-peer (P2P) networks has drawn great interest, since it can not only accommodate large numbers of users, but also handle peer heterogeneity. However, there’s still a lack of comprehensive studies on chunk scheduling for the smooth playout of layered streams in P2P networks. In these situations, a playout smoothing mechanism can be used to ensure the uniform delivery of the layered stream. This can be achieved by reducing the quality changes that the stream undergoes when adapting to changing network conditions. This paper complements previous efforts in throughput maximization and delay minimization for P2P streaming by considering the consequences of playout smoothing on the scheduling mechanisms for stream layer acquisition. The two main problems to be considered when designing a playout smoothing mechanism for P2P streaming are the fluctuation in available bandwidth between peers and the unreliability of user-contributed resources—particularly peer churn. Since the consideration of these two factors in the selection and scheduling of stream layers is crucial to maintain smooth stream playout, the main objective of our smoothing mechanism becomes the determination of how many layers to request from which peers, and in which order. In this work, we propose a playout smoothing mechanism for layered P2P streaming. The proposed mechanism relies on a novel scheduling algorithm that enables each peer to select appropriate stream layers, along with appropriate peers to provide them. In addition to playout smoothing, the presented mechanism also makes efficient use of network resources and provides high system throughput. An evaluation of the performance of the mechanism demonstrates that the proposed mechanism provides a significant improvement in the received video quality in terms of lowering the number of layer changes and useless chunks while improving bandwidth utilization.     

P2P流媒体数据调度研究综述

P2P流媒体通过利用网络上普通主机节点的资源来提供流媒体数据服务,是一种扩展性好、性价比高的流媒体服务体系。数据调度是P2P流媒体研究中的核心问题,流媒体中严格的服务质量要求、Peer节点状态的不稳定性以及其带宽资源的有限性是其面临的主要挑战。介绍了近几年来该领域基于单播树、多组播树和随机拓扑三类典型的数据调度策略的原理特点和Peer节点搜索定位技术的研究进展,指出了未来的几种研究方向。     

Towards efficient video chunk dissemination in peer-to-peer live streaming

There are substantial differences in chunk dissemination manner between P2P live streaming and BitTorrent, and inappropriate algorithms will result in inefficiency of live streaming systems. In this paper, we study the chunk dissemination of P2P live streaming, and introduce a discrete and slotted mathematical model to analyze chunk selection algorithms, including rarest first algorithm and greedy algorithm. Moreover, we present a performance metric to evaluate chunk selection algorithms, as well as the optimization function for the exploration of chunk dissemination strategies. We point out the causes of poor performance of these algorithms, and propose a service request randomization mechanism to promote the use of peer resources, which can prevent chunk requests from rendezvous on a few of peers. Simultaneously, we employ weight assignment strategies to avoid excessive requests for rare chunks. Besides, we present an enhanced model, which adds node degree constraint, to improve our model. We revisit the chunk selection algorithms based on the enhanced model. The results of simulation experiments validate our theoretical analysis and indicate that the weighted randomization mechanism is resilient to flash crowd and peer churn, and can improve the performance of P2P live streaming.     

Adaptive content-and-deadline aware chunk scheduling in mesh-based P2P video streaming

In mesh-based Peer-to-Peer (P2P) live video streaming systems packet scheduling is an important factor in overall video playback quality. In mesh based P2P video streaming systems, each video sequence is divided into chunks, which are then distributed by multiple suppliers to the receivers. The suppliers need to be coordinated by the receiver through specifying a transmission schedule for each of them. Many previous studies on scheduling of P2P streaming tend to mainly focus on networking issues which strongly depend on a particular P2P architecture such as tree or mesh. These algorithms suffer from some design issues: 1) they are too complex to deploy, 2) they do not take video characteristics into account and 3) they do not have sender-side transmission policy. To address all three of these problems, we propose a new chunk scheduling scheme which consists of two parts: i) receiver-side scheduler and ii) sender-side transmission order scheme. The proposed receiver-side scheduler considers the contribution level of each video frame as well as the frame’s urgency in order to define a priority for each video frame. It attempts to request frames with highest priority from peers which can deliver them in a shorter time. We also design a new chunk transmission order scheme that decides which requested chunk will be sent out first based on its importance to the requesting neighbor. Our simulation results show that the proposed scheduling scheme improves the overall quality of the perceived video in mesh-based P2P video streaming architectures substantially.     

P2P流媒体直播分布式缓存替换算法研究

P2P流媒体直播系统中分布式节点缓存区别于传统的Client/Server缓存结构,节点的实时同步给缓存管理提出较大挑战。分析了分布式缓存空间利用率的决定因素,通过节点成功请求比率,缓存的fresh度及数据分片点击率3个指标来评估节点缓存空间利用率,提出了频度限制与改进的LRU相结合的K-Degree&LRU2缓存替换算法。仿真实验结果表明,该算法较传统的FIFO、LRU算法具有更高的执行效率。