Efficient Streaming Proxy Server Caching Technology

Streaming media technique includes the data transmitting technique as well as streaming media data coding technique. Accessing multimedia files with streaming media technique will generate data streams with long time existed and big bandwidth consumed. Streaming media cache in the proxy server do good in reducing network traffic of streaming media.The continuous data transmission brings great challenge to the proxy sever, and it is of great importance to find out how to reduce cache-writing load in order to avoid the damage to the proxy service.Cache method is the core part of the streaming media cache, since it determines how proxy server works. The extant cache method is only aiming to reduce the amount of data transmission and the delay of getting access to the media. However, it ignores the load increase while cache writing happens. Such ignorance will reduce the real efficiency of the cache method. Even the Adaptive&Lazy method is subjected to such inefficiency, which is the best streaming media cache method for reducing network traffic.The current dissertation is primarily concerned with the new streaming media cache method which has good effect on both cache-writing efficiency and reducing network traffic. We propose that the tradition mode separated cache selection from replication selection should be changed. We do a lot of research work on the sum of cache-writing load and the real-time cache-writing load. The following are key points in the paper:(1) A new admission model of cache-writing is proposed to research the writing process of the segmented cache. Based on the observant that release action can reflect increasing of request frequency, we define a new concept"request zone". We suggest using the continuous request to represent access frequency of media object, and design the intermissive write admission model (IWAM). It is proved that IMAM has no any obvious blemish and does better than the Lazy model both on transmission saving and writing efficiency.(2) We design the new cache method based on the IWAM that will save the amount of transmission as well as writing efficiency. More specially, we propose: (a) a new admitting policy based on IWAM, (b) some new factors of replication, (c) a new release method. Through simulation by actual access log, the new cache method has the same transmission saving and promote the writing efficiency as three times as the Adaptive&Lazy method.(3) Based on the phenomenon that the peak period of cache-writing load is the same as the customer access, we propose that bandwidth judgment of cache-writing can be used to limit biggest cache-writing load. We explain how the definition and judgment of cover scope in the cache-writing can impact the bandwidth usage judgment. Experiments result revealed that the judgment have promoted the test result in cache-writing efficiency greatly.(4) In order to support test using large scale access log, we also developed the LittleDuck simulator. Comparing to the MiddleSim simulator, Littleduck shortened the simulating time to 1/60 or less with the .4% outcome variance. Experiments that simulating the process with load limitation were conducted to test the performance. Comparing to the Adaptive&Lazy method, the new method's improvement in transmission saving is 10%...