Video-on-Demand Broadcasting Protocols

Introduction

Video-on-demand (VOD) will one day allow clients to watch the video of their choice at the time of their choice. It will be less complicated to use than a typical web browser, and it will only require that the clients have a set-top box (STB) connected to their television set. The STB will allow the clients to navigate a VOD server's video library, and then it will handle the reception and display of the video once the clients make a selection. It is possible that the VOD service will even be interactive, allowing clients to use VCR controls such as pause, rewind, and fast forward while viewing their requests. In this case the STB will also be responsible for communicating the desired interactions to the VOD server.

Unfortunately, video is an expensive medium to distribute. Even using a compression scheme such as MPEG-2, a stream of video data will take a bandwidth of at least 4 Megabits per second (Mb/s)—or roughly 70 times the capacity of a 56K modem. This sort of bandwidth will not necessarily be a problem for the client, but the VOD server will be limited by its bandwidth capacity, and so it must be able to service as many clients as possible in the bandwidth it has available.

Simply dedicating a unique stream of video data to each client request is not a good idea because that can quickly use up all of the available bandwidth on the VOD server. Better, more efficient strategies are needed.

One such set of strategies involves broadcasting videos to clients. Video rental patterns suggest that most client requests will be for the most popular 10–20 videos, and so broadcasting these videos ensures bounds on the amount of time clients must wait before they can watch their requests and on the total amount of bandwidth each video will require.

In the remainder of this chapter we will first define some common terms and concepts used by broadcasting protocols, and then we will describe the protocols themselves.