A novel scheduling approach to concurrent multipath transmission of high definition video in overlay networks

摘要

Recent advancements in network infrastructures provide increased opportunities to support video delivery over multiple communication paths. However, the high definition (HD) video transmissions still pose crucial challenges due to the high throughput demands and large-size video frames. Motivated by optimizing the delay performance for concurrent multipath transmission of HD video, we propose a novel scheduling approach dubbed FSWG (Frame Splitting based on Weibull distribution and Graph theory) that aims to minimize the end-to-end frame delay while alleviating out-of-order arrivals. First, we analytically construct a delay performance model for HD video streaming in multipath overlay networks based on Weibull distribution and graph theory. Second, we formulate the frame splitting over parallel paths as a constrained optimization problem of minimizing total frame delay and derive its solution based on the water filling algorithm. Third, we design a multipath video transmission system to implement the proposed scheduling approach. The performance evaluation is conducted through extensive simulations in QualNet using H.264 video streaming. Experimental results show that FSWG outperforms the existing schemes in terms of Mean Opinion Score (MOS), Peak Signal-to-Noise Ratio (PSNR), and delay performance metrics.