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.
