Study the Effect of FEC on Video Streaming over the Networks

Abstract: Context: Video streaming over best-effort networks is complicated by a number of factors including unknown and time-varying bandwidth, transmission delay, and high packet losses. Excessive packet loss can causes significant degradation in quality of video perceived by users of real-time video applications. Recent studies suggest that error control schemes using Forward Error Correction (FEC) is a good candidate for decreasing the negative impact of packet loss on video quality. With FEC scheme, redundant information is transmitted along with the original information so that the lost original video data can be recovered at least in part from the redundant information. Although sending additional redundancy increases the probability of recovering lost packets, it also increases the bandwidth requirements which in turn increase loss rate of the video stream due to congestion. For this reason FEC is characterized as effective when the redundancy is sufficient to recover the lost data, ineffective when the redundancy is little to recover the lost data and inefficient when the redundancy is high for the lost data. To provide best performance for the streaming application and the network, it is important to determine, in real time, the proper amount of redundant information according to the loss behavior of the network. However it not clears how to choose optimal redundancy given the constraints mentioned above at any given point in time. Objective: This thesis investigated the limitations of existing static FEC schemes and suggested alternative approaches that can achieve effectiveness through sending the appropriate amount of FEC redundancy in real-time. Three dynamic redundancy adaptation control systems On-Off, Proportional, Proportional-Integral have been design and implemented based on existing conventional FEC mechanisms. Moreover the performance evaluation of each mechanism is performed on simulated environment. Method: A performance evaluation methodology using network simulation and key evaluation criteria to test the mechanisms under different network conditions and the scenarios for each different network are considered. The proposed mechanisms were implemented in a simulation environment by using NS2. After the implementation and validation of the techniques, several simulation experiments have been conducted to study the performance of each redundancy adjustment schemes. Results: The result of the simulation experiments and performance analysis showed that both Proportional and Proportional-Integral (PI) redundancy controller based on Adaptive Forward Error Correction (AFEC) mechanisms significantly reduce the number of discarded video blocks as compared with the On-Off based scheme. On the other hand On-Off based AFEC mechanisms added small number of excess redundancy. The amount of excess redundancy added by the Proportional and PI controllers significantly increases as the packet drop rate of the transmission channel increases. Conclusion: Based on the findings of this study, using the PI adaptation based AFEC mechanisms is potentially viable option for enhancing the performance of video streaming applications over the network.