QVOD点播系统的测量研究

Study on Measurement of the QVOD VoD System

针对大规模P2P点播系统监管难度大、网络利用率低的问题,研究了流行的QVOD点播系统。采用逆向工程技术分析了QVOD点播系统的系统架构和节点分发协议,提出了一种高效的爬行停止条件模型,根据该模型实现了一款分布式爬行器QVOD_Crawler,使用该爬行器对QVOD点播系统进行了为期2周的主动测量,对节目的在线人数、节点的动态性等进行了研究。主要发现有:QVOD点播系统中节目的在线人数在时间上具有以天为单位的自重复性,且与人们的作息时间密切相关;QVOD点播系统中同一个节目的加入节点和离开节点数量基本持平,且加入与离开的节点数和节目在线人数有同样的趋势;QVOD点播系统中用户的耐心比PPLive直播、点播系统中用户的耐心强,QVOD节点在线时长小于1 000 min时属于幂率分布。研究结果可为大规模点播系统的监测和系统优化提供重要依据和基础。

In order to deal with the management difficulty and low network availability of large-scale P2P VoD systems,a kind of popular P2P VoD system named QVOD was investigated. The architecture and peer-distributing protocol of the system were analyzed with reverse engineering technology. A model was proposed to determine the stop crawling condition. According to the model,a distributed crawler named QVOD_Crawler was developed. Active measurements were used in the experiment. The program size and the churn of the system were studied. The results showed that the number of the online users in the QVOD system remained with a constant daily viewing pattern. The churn ratio had the same trend with the number of online users. The peers in the QVOD systems were more patient than those in the PPLive VoD system and other P2P live streaming systems,and the peer-life time followed a power-law distribution when the peer-life time was smaller than 1 000 minutes. The experiment results could be used as an important basis for the management and optimization of large-scale P2P VoD systems.