Loading...
Please wait, while we are loading the content...
Similar Documents
Chapter 2 Error Control for Interactive Continuous Media Applications 2.1. Requirements of Continuous Media Applications 2.2. Avoiding Losses 2.2.1. Forward Error Correction (fec) 2.2.2. Concealment 2.3. Making Retransmission Work 2.3.1. Playout Buffering 2.3.2. Gap-based Loss Detection
| Content Provider | Semantic Scholar |
|---|---|
| Abstract | In the previous chapter we characterized continuous media (CM) applications as applications that have very strict timing requirements. We gave examples like audio and video applications, which require periodic transmission of data (e.g., 30 frames per second for MPEG video), and pointed out that the utility of such applications depends on the regular and timely delivery of data in order to ensure a smooth presentation, free of annoying artifacts like image distortion and jumping. While the strict timing requirements of CM applications make error control difficult, we also pointed out that compression schemes like MPEG make error recovery even more important because loss of a compressed data packet typically results in more information loss than loss of an uncompressed packet. We made a distinction between interactive and stored media applications, pointing out that error control in interactive CM applications is harder to implement than stored media applications, because in the former data needs to be transmitted almost immediately after generated to preserve the feeling of interaction. Long round-trip-times (RTT), like those experienced in coast-to-coast communication, complicate error control even further because they limit the time available for recovery. At the time our work was published[85], it was commonly believed that retransmission-based error recovery was inappropriate for interactive continuous media (CM) applications, because of latency. However, we felt that retransmission was still an attractive option, because it requires minimal network bandwidth and processing cost, and argued that despite its latency, retransmission can be adapted for use even in interactive CM applications. 16 In this chapter we present the design and implementation of a retransmission-based error control scheme for CM applications, which aims to provide the best possible reliability at a minimal cost, without violating the application's timing constraints. To achieve this goal, we have enhanced selective-repeat retransmission with the following mechanisms: (1) playout buffering to increase the time available for recovery; (2) gap-based rather than timer-based loss detection to minimize loss detection latency; (3) implicit expiration of sender retransmission buffers to eliminate acknowledgments ; (4) conditional retransmission requests to avoid triggering late, unnecessary retransmis-sions; and (5) data integrity information delivery to the application to aid in concealment. Our experiments have shown that the mechanism significantly reduces observed loss without violating the application's delay constraints. For example, it has reduced observed loss by orders of magnitude for both random and bursty loss. We have implemented this protocol in the kernel of NetBSD … |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.cs.colostate.edu/~christos/thesis/chp2.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Chapter |
