视频会议系统在互联网上得到了广泛的、成功的应用,标清、高清视频会议也是研究的新热点。然而,AccessGrid、数字视频传输系统(digital video transport system,DVTS)、VideoLAN客户端(VideoLAN Client,VLC)和Ultra Grid等各种视频会议系统大都面临着可扩展性的问题。带宽成为限制系统容量的主要因素。该文提出了一种通用的可扩展分布式标清视频会议系统的结构以及转发模型设计,通过采用多台数据转发服务器进行协作转发,同时在控制平面和数据平面进行了结构、系统组件行为及转发模型设计,并对负载进行合理分配和均衡。在DVTS的基础上,实现了上述的系统结构及转发模型设计,在中国教育科研网(CERNET)上实现了基于DV的多点对多点标清视频会议系统DVTS Plus,并进行了现网实验验证。实验结果证明:分布式的DVTS Plus视频会议系统可以支持更大的系统容量和更方便的部署,并具有良好的可扩展性。
在无线网络中为了节能,主流做法是将AP(Access Point,无线接入点)进行聚簇,在网络闲时关掉一些AP而让剩下的AP来支持用户.聚簇的关键在于合理的聚簇依据.相关研究中通常利用AP之间相互监听到的RSSI(Received Signal Strength Indication,接收信号强度指示)进行聚簇.AP之间的RSSI并不是一个合理的依据,因为AP之间是否可以相互替代应从用户的角度出发,而不应从AP角度来衡量.此外,相关研究中还缺乏覆盖损失的量化描述,以及应对用户数目震荡性变化的措施.本文利用AP监听到的用户的RSSI构建了AP可替代性的数学模型,以期为AP聚簇提供合理的依据,并基于AP可替代性设计了节能机制.分析表明AP可替代性模型是有效、稳定的.利用AP可替代性模型可以量化描述AP可替代性与覆盖损失之间的关系,从而可以根据实际需求来确定聚簇的粒度,实现了节能与保持覆盖的两相兼顾,体现了本文机制的灵活性.此外,本文还设计了应对用户数目震荡性变化的双阈值策略,有效的避免了聚簇算法被反复无意义的执行.实验结果表明,相比于同类研究,本文的节能机制在损失很小的节能效果的同时,显著地减小了覆盖损失.
As a path vector protocol, Border Gateway Protocol （BGP） messages contain an entire Autonomous System （AS） path to each destination for breaking arbitrary long AS path loops. However, after observing the global routing data from RouteViews, we find that BGP AS Path Looping （BAPL） behavior does occur and in fact can lead to multi-AS forwarding loops in both IPv4 and IPv6. The number and ratio of BAPLs in IPv4 and IPv6 on a daily basis from August 1,2011 to August 31, 2015 are analyzed. Moreover, the distribution of BAPLs among duration and loop length in IPv4 and IPv6 are also studied. Several possible explanations for BAPL are discussed in this paper. Private AS Number Leaking （PANL） has contributed to 0.20% of BAPLs in IPv4, and at least 1.76% of BAPLs in IPv4 were attributed to faulty configurations and malicious attacks. Valid explanations, including networks of multinational companies, preventing particular AS from accepting routes, also can lead to BAPLs. Motivated by the large number of PANLs that contribute to BAPLs, we also study the number and the ratio of PANLs per day in the 1492 days. The distribution of the private AS numbers in all of the PANLs is concentrated, and most of them are located in the source of the AS paths. The majority of BAPLs resulted from PANLs endure less than one day, and the number of BAPLs which are caused by two or more leaked private ASes are much larger than that of BAPLs which are caused by one leaked private AS. We explain for this phenomenon and give some advices for the operators of ASes.
