Research On Proxy-based High Efficient Caching Technology For Streaming Media

Streaming Media Proxy caching method can reduce the network transmission capacity of streaming media effectively. However, because the continuous data stream will bring great challenges to the proxy caching, it is an important research problem that how to guarantee the service performance with the lowest disk and bandwidth consumption in the situation of ensuring high service quality. Nowadays, the mainstream cache replacement algorithms based on the popularity generally lack quantitative analysis and estimate for the content popularity. When abstracting segmentation factor, the mainstream segment-based algorithms also lack the quantitative analysis and consideration for the client behavior character. While the delicate granularity cache algorithm required huge additional storage space, it has very low value in practice. Aimed at the problems above, after investigating the streaming media proxy cache technology, content popularity and segment-based adaptability streaming media caching technology based on the quantitative analysis of the content popularity was proposed in this paper.In this paper, by the collection of actual log files, K-transformed Zipf-like model was introduced for analysis of client behavior character. In the quantitative analysis of internal popularity, two-point popularity idea was introduced so that internal popularity disturbances could be prevented , precise contents popularity collection and estimates strategies were also proposed. During the media segmentation, media segmentation strategy based on segmentation factorφwas proposed. On the basis of the segmentation strategy, media block prefetching strategy, cache access strategy and cache value estimating functions based on the content popularity were provided. According to these functions, replacement policy and release strategy were also suggested to form a complete streaming media caching mechanism.A proxyer simulator was developed independently. Related simulations of popularity and segment-based adaptability caching algorithms, exponential segment-based caching algorithm, and LRU isometric segment-based caching algorithm have been conducted on it. The byte hit rate and the backbone bandwidth consumption savings rate were selected as the algorithm performance evaluation index. The results suggested that with the same network configuration, algorithm proposed in this paper got superior algorithm performance than LRU isometric segment-based algorithm and exponential segment-based caching algorithm.Based on the research of the proxy based high-performance streaming media caching technology, K-transformed Zipf-like model was introduced to describe the distribution characteristics of streaming media content popularity in this paper. Two-point popularity idea was introduced, and contents popularity was estimated quantitatively. In the segmentation strategy, precise contents popularity estimates strategies of K-transformed Zipf-like model were set as the primary factor while abstracting the segmentation factorφ. A segmentation mechanism for the distribution character of media popularity based on the segmentation strategy was proposed so that an integrated streaming media caching mechanisms including prefetch, access, replace, and release strategies were formed. In this paper, based on popularity and segment-based adaptive streaming media caching mechanism, a good balance of segment-based caching algorithm in the cache utilization and accuracy of algorithms were created, while improving the streaming media proxy server's cache efficiency.