Method, equipment and system for implementing stream media caching replacement

Abstract

The invention discloses a realizing method of stream media buffer transposition, which comprises the following steps: a request from a client side node is received, and the stream media object required by the client side node is obtained; whether the size of the stream media object required by the client side node is larger than current buffer space is judged, if the judgment result is yes, the transpositional factor of each stream media object stored in the buffer depending on the network delay, duplicate amount, size and popularity of the stream media object stored in the buffer is obtained, and the size of the transpositional factor corresponds to the size of the priority level that the stream media object should be transposed; the stream media object of the larger priority level in the buffer is transposed with the obtained stream media object. Meanwhile, the invention discloses a device and the system for realizing the stream media buffer transposition. The invention has the advantages of considering synthetically the factors affecting the buffer effect, optimizing the main performance indexes of the stream media server, and improving the utilization ratio of the resource.

Description

technical field [0001] The invention relates to peer-to-peer network streaming media technology, in particular to a method, device and system for realizing streaming media cache replacement. Background technique [0002] The application of peer-to-peer network (P2P, Peer to Peer) streaming media is developed on the basis of P2P file exchange. P2P file exchange is that users download the required files (such as video streaming media) from the server. After that, the user can use the file, such as watching the downloaded video streaming. And the application of P2P streaming media technology can make the user watch the video streaming media (which may not be a complete video streaming media), and share resources with other users who watch the video streaming media. Other users watching the video streaming media download the video resource at the same time, realizing playing while downloading. Therefore, compared with the P2P file exchange technology, its resource utilization ...

AI Technical Summary

Problems solved by technology

However, this algorithm has the problem of cache pollution (Cache Pollution): that is, when a popular streaming object becomes no longer popular, the streaming object will still remain in the cache for a long time, thus preventing other new ones. popular streaming object to replace it

By displacing the largest streaming object in the cache, it increases the chance of small streaming objects being cached, thus improving the streaming object hit rate, but affecting the byte hit rate

In addition, another disadvantage of this algorithm is that some cached streaming media objects are hardly accessed or rarely accessed, but occupy the cache space of the system, and also have the problem of "cache pollution"

[0055] (1) The size of a typical Web object is usually only a few K to tens of K bytes, so the basic unit of Web cache replacement is generally a complete object; while a streaming media object is usually much larger than a Web object, so if the entire Streaming media objects are obviously inappropriate as the basic unit of cache replacement, so it is necessary to improve the Web cache replacement algorithm before it can be applied to streaming media

[0056] (2) Most of the current streaming media objects on the Internet are static streaming media objects, which usually have the nature of write once and read many times (WORM). When a static streaming media object is placed on the Internet, its WORM nature is rarely changed

[0066] (1) The traditional cache replacement algorithm is used to improve the performance optimization index of the cache hit rate. For the streaming media server, the media quality received by the user, the user's startup delay and the consumption of multimedia data on the network are the main indicators of the streaming media server. Therefore, the traditional The cache replacement algorithm is difficult to meet the requirements of streaming media servers

[0067] (2) The current streaming media cache replacement algorithm only considers a certain factor that affects the cache effect, such as popularity, transmission cost, cache gain, etc., but does not consider these factors comprehensively to formulate a feasible cache replacement algorithm

[0068] (3) At present, most cache replacement algorithms are not very comprehensive for the calculation of the popularity of streaming media. For example, some calculations are simply based on the number of historical clicks. Therefore, if the streaming media object is no longer popular after a period of time, but Due to its high historical click rate, it will be cached in the cache for a long time before being replaced, or even not replaced, occupying system resources and resulting in low system resource utilization

[0069] (4) At present, most of the cache replacement algorithms are aimed at the traditional C / S structure streaming media system, and there is no cache replacement algorithm for the hybrid P2P streaming media system

Images