New conference paper: Application Layer Throughput Control for Video Streaming over HTTP2

Our latest research on throughput control for video streaming has been accepted for presentation at Eighth International Conference on Communications and Electronics (ICCE).  In this work, we tackle the problem of throughput control for multiple streaming clients at application layer using HTTP2 flow control feature. The proposed framework includes a throughput estimation scheme and an application layer based throughput control scheme which utilizes HTTP2’s Flow Control mechanism. Experiment results show that the proposed framework can effectively control throughput of individual clients.

Why do we need "throughput control"?

In practice, there are many cases where multiple videos are concurrently streamed over a same network connection with limited bandwidth.  A common example is when people in a same local network (e.g., home, office) watch online videos at the same time. Another example is camera surveillance system where a large number of video feeds from surveillance cameras is streamed to a processing data. In addition, those video streams may have different characteristic (e.g., bitrates) or have different levels of importance. Thus, it is desirable to be able to allocate available network bandwidth in an efficient manner. For instance, in case of camera surveillance system, video streams from important cameras should have higher throughput than those from less important ones.  

Why application layer matter?

In our previous work [1], we successfully control throughput of streaming client by adjusting TCP flow control window size. However, it may not always possible to perform such control in practice, such as in browser-based streaming, or in mobile devices. Thus, it is desirable to be able to control throughput at application layer.

[1] Duc V. Nguyen, Huyen Tran, Pham Ngoc Nam, Truong Cong Thang, "Cross-Layer Management for Multiple Adaptive Streaming Clients in Wireless Home Networks", IEICE Transactions on Information and Systems, March. 2018.