Before Streaming Middleware To The Research And Implementation Of Error Correction Strategy

With the rapid innovation of networking and streaming media technology, and the steady increase in broadband deployment, IP (Internet Protocol) based streaming applications have gradually become main stream applications, which occupy a leading position in the allocation of networking bandwidth resource. Digital TVs are the most common and critical form of applications among streaming applications. Recently, there is a trend of nationwide rapid promotions of IPTV and satellite based Digital TV. In the streaming media transmission, the video bit rate increases along with the increasing demand of users, and coupled with the increasing complexity of the overall network environment, that makes the original IP multicast very difficult to meet the currently large-scale, high-quality service demand. Restricted by the best effort feature of IP network, packet loss is widespread among video streaming service. Thus, IP network packet loss is the bottleneck of improving Quality of Service (QoS) of existing streaming services. Thus optimization and improvement are needed. Although the industry has put much effort in research and development of streaming applications and services, these products mostly come short in identical functionality, poor compatibility, low scalability and other shortcomings, which greatly reduced their market value. To this end, there is a trend to introduce a unified and compatible development environment to provide a common interface and communication protocol functions, which enables mutual collaboration and overall service quality improvement.Based on the above two requirements and the existing research results, we first analyze, design and implement an encapsulation for the functionality of each module in streaming reliable transmission based on FEC(Forward Error Correction) packet recovery, and merge it into the streaming application middleware. It provides a unified, compatible, interoperable development environment for different developers. Then, for the need of transmitting high bit-rate video in a complex network environment, this paper proposes a highly optimized RS(Reed-Solomon) FEC strategy with different principles. Also, to further assess the effect of our FEC strategies, we use a variety of media sources in a number of experiments, and do a detailed analysis on quality of recovery of transmitting different bit rate sources comparing to standard FEC strategy.The assessment results showed that our optimized FEC strategy not only can guarantee a reliable transmission of media data to the client, enable normal display, but also can greatly control the volatility of the CPU/memory usage on server side and essentially reduce the decoding complexity/time on client side, which show promise for the actual deployment in future.