基于对等网络的VoIP 关键技术研究
文献类型：学位论文
| 作者 | 张建东 |
| 学位类别 | 博士 |
| 答辩日期 | 2007-05-25 |
| 授予单位 | 中国科学院声学研究所 |
| 授予地点 | 声学研究所 |
| 关键词 | VoIP 对等网络 服务质量 多路传输 自适应缓存播放 负载平衡 |
| 其他题名 | Studies on Key Technologies of VoIP over Peer-to-Peer Network |
| 学位专业 | 信号与信息处理 |
| 中文摘要 | 近年来，VoIP（Voice over Internet Protocol）技术得到了广泛的应用。但现有VoIP应用大多采用客户端/服务器架构实现，配置管理较复杂，鲁棒性和可扩展性稍差，在部分场景下无法应用。基于对等网络实现VoIP，则可以提高鲁棒性和可扩展性，具有更高的性价比，并且可以做到零配置。 本文针对基于对等网络的VoIP关键技术进行深入研究，主要包括用于VoIP的对等网络架构、对等网络中VoIP信令路由机制、语音分组传输控制和播放方法以及对等终端间注册用户负载平衡方法等。经过上述研究，本文针对信令传输质量、语音质量和注册用户负载平衡等问题提出相关改进方案，提高基于对等网络的VoIP服务质量和终端实现的性价比。 本文的主要研究成果如下： 1、提出一种基于对等网络的VoIP应用架构。该架构基于Chord对等网络，并针对VoIP应用需求对节点加入、离开、维护等操作进行了改进。 2、提出一种对等网络中VoIP信令的低延迟路由方法。该方法基于查询附加随机采样方法，使用低延迟的节点组成指针表，采用递归查询和即时回应相结合的方法，降低信令路由单跳延迟和失效节点的影响。仿真表明，与原有Chord路由方法相比，该方法有效的降低了VoIP信令路由的平均延迟。 3、提出一种基于对等网络的语音分组多路传输方法。该方法根据低延迟指针表中节点的信息选择中继节点传输冗余的语音分组，提高中继路径的有效性，保证语音分组的传输质量。 4、提出一种基于统计和服务质量的多路自适应缓存播放算法。该算法估计多路语音延迟分布函数，根据延迟和丢包率的关系及服务质量目标函数估计分组播放时间。仿真表明，与单路自适应播放算法相比，语音质量有所提高。 5、提出一种注册用户负载平衡方法。该方法基于非固定的节点标识，改进终端节点的加入过程，采用两级阈值方法监控整个网络的负载状况，使用局部平均和范围传递方法转移注册用户负载。仿真表明，相比原有Chord网络，使用该方法可以有效降低最大注册用户负载量和方差。 |
| 英文摘要 | The research on VoIP has gained much attention in recent years. However, because the majority of the existing approaches to deploy VoIP services adopt the client/server mode, which have some inherent drawbacks, including complexity of network configuration and management, lack of robustness and poor scalability, VoIP services can not be accomplished under many certain application environments. A peer-to-peer (P2P) network based VoIP will afford more robustness and scalability, and can be implemented with zero configurations and cost-efficiency. This thesis gives further studies on key technologies of VoIP over P2P network, which mainly focus on the P2P overlay network utilized by VoIP service, the VoIP signaling routing mechanism over P2P network, voice packet transporting control and playout scheme, registered user load balancing scheme among peer nodes, and etc. Following from the above studies, this thesis proposes some novel and effective approaches to address the problems including signaling transporting quality, voice quality and registered user load balance, which improve the quality-of-service (QoS) of P2P-based VoIP, and afford more cost-efficiency for the embedded realization. The main research contributions of this thesis include: 1. A P2P-based architecture for VoIP application is proposed. This architecture adopts Chord protocol, and improves the operations of node joining, leaving and stabilizing in order to satisfy the requirements of VoIP. 2. A low delay VoIP signaling routing scheme over P2P network is proposed. Based on lookup-parasitic random sample method, the scheme uses low delay finger table to reduce one hop delay of the forwarding signaling and decrease the affect of node failures with the recursive and instant-feedback mechanism. Our simulation shows that, compared with the original Chord routing scheme, the scheme reduces the average delay of signaling routing effectively. 3. A multi-path packets transporting scheme is proposed. Based on a P2P network, the scheme selects relaying nodes for redundant packets transporting according to a low delay finger table. It improves the efficiency of the relay path, and guarantees the transporting quality of voice packets. 4. A statistical and QoS based multi-path adaptive playout buffering algorithm is proposed. The algorithm estimates a distribution function of the multi-path voice packets via the received packet delaying information, and calculates the packet playout delay according to the relation of delay and lost ratio by maximizing the QoS target function. Our simulation shows that, compared with the uni-path adaptive playout buffering algorithm, the algorithm improves voice quality. 5. A scheme of load balancing of the registered users is proposed. Based on unfixed node identification, the scheme improves the operation of node joining, monitors the registered user loads with two-level thresholds, and transfers user loads by local-averaging and range-transferring method accordingly. Our simulation shows that, compared with the original Chord protocol, the maximum and variance of registered user loads are effectively reduced by the scheme. |
| 语种 | 中文 |
| 公开日期 | 2011-05-07 |
| 页码 | 149 |
| 源URL | [http://159.226.59.140/handle/311008/84]  |
| 专题 | 声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文 |
| 推荐引用方式 GB/T 7714 | 张建东. 基于对等网络的VoIP 关键技术研究[D]. 声学研究所. 中国科学院声学研究所. 2007. |
入库方式： OAI收割
来源：声学研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。