4.2. A Taxonomy of Deadline-driven Approaches

We can straightforwardly generalize double buffering to N buffering by using N buffers and prefetching N-1 blocks before initiating a display. The system can continue to request a block every time a memory buffer becomes empty. This reduces the probability of hiccup to p[ω > (N − 1)Tp] because the retrieval time of a block must now exceed (N-1) time periods in order for a display to incur a hiccup. Assuming a clip consists of n blocks which have display sequence from B0 to Bn−1 and N buffers, at most N-1 blocks[1] (from B0 to BN−2) can be prefetched and accumulated before starting display to provide a more tolerable variance in block retrieval time for the consecutive blocks (BN−1 to Bn−1). This is ...

Get Streaming Media Server Design now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.