Cache architecture for on-demand streaming on the Web
ACM Transactions on the Web (TWEB)
| Hi-index | 0.00 |
Since multimedia data are characterized by continuity and massive volume, compression techniques such as MPEG are used for efficient transmission. However, a transmission error occurring at a frame tends to be propagated to other frames because MPEG frames crossreference one another upon presentation. Due to time constraints, it is usual to discard without retransmission the data which have transmission errors or which cause serious delay. But in cases of multimedia applications where retransmission is required (for example, missioncritical applications), the QoS may be severely deteriorated by the continuous loss caused by propagation if damaged frames are simply discarded. In this paper, we propose a selective retransmission method where the number of retransmissions can be reduced without losing QoS and data continuity. Since the proposed method keeps track of error propagation ranges using virtual buffers which have information about the sizes and types of frames before transmission, the number of retransmissions from the transmitter to the receiver can be minimized by selecting and retransmitting the recoverable data only. Simulation results show the proposed method reduces retransmission cost considerably while maintaining QoS similar to the case of unconditionally retransmitting all the damaged data.