Skip to Main Content
As broadband networks using Fiber-to-the-x (FTTx) technologies are being increasingly deployed in access networks, video service, especially VoD (Video-on-Demand), is becoming more attractive to deploy. To provide efficient VoD service, a cost-effective service model is very important and a lot of research has been conducted over the past decade. This paper reviews the existing literature on this topic, focusing on the user behavior in VoD services and bandwidth-saving multicast streaming schemes, which are the most important aspects of VoD service. First, we review the user behavior in VoD such as video popularity, daily access pattern, and interactive VCR (Videocassette Recorder) properties from recent data. Each video title's rental frequency, i.e., video popularity, follows the Zipf distribution, and this popularity can change with time or by service provider's recommendation of videos. This overall request frequency for each video constitutes a specific pattern throughout the day and has a similar pattern every day. Second, we review the bandwidth-saving streaming schemes such as broadcasting, batching, patching, and merging, which use multicast streaming technologies and user buffer memory. We review the mechanism of each multicast streaming technology and compare their differences. We also review the recent trends on multicast streaming technologies, which is summarized as hybrid architecture which combines several multicast streaming technologies to obtain better performance. Next, we review how these multicast streaming technologies implement interactive VCR functions. We classify the VCR interactivity into discontinuous and continuous VCR actions and examine the principles for VCR support in multicast streaming schemes: caching some video data for discontinuous VCR support and allocating contingency channels for continuous VCR support. We review mechanisms of VCR support for different multicast streaming schemes. Through this survey, we provide an in-de- th understanding of VoD service deployment.