HLS content source returning and caching realization method

Abstract

The invention provides an HLS content source returning and caching realization method, and relates to the technical field of internet multimedia. In the HLS content source returning and caching realization method, after a caching server receives an HLS content request sent by a client-terminal, a session object is allocated for the client-terminal and unique identification is generated; if the HLS content is locally cached, current and subsequent requests are directly responded, and if the HLS content is not cached, the caching server requests to a source station for all the index files of the HLS content, and addresses are modified; and after media fragment requests sent by the client-terminal each time are received, the caching server requests to the corresponding source station to return the media fragment of the time to the client-terminal and continuously store the media fragment of the time in a large file according to the sequence of the index files until all the media fragments are stored. The client-terminal behavior can be completely tracked, adaptability of the source returning process is high, efficiency of butting with the source station is high, and storage and utilization efficiency and disk throughput are high.

Description

technical field [0001] The invention relates to the field of Internet multimedia technologies, and specifically relates to a method for realizing back-to-source and caching of HLS content. Background technique [0002] HLS (HTTP Live Streaming) is an HTTP-based multi-bit rate adaptive streaming media specification proposed by Apple. At present, HLS has become the mainstream streaming media playback specification on the Internet and mobile Internet. This specification supports a content with multiple media files with different bit rates, and each bit rate media file is divided into many small files with a duration of about ten seconds. A content has several different bit rates, and which small files a bit rate consists of. These information are organized in several index files. These index files are in M3U8 format, so they are also called M3U8 files. [0003] The process of the client playing HLS content is essentially the process of downloading index files and fragment fil...

AI Technical Summary

Problems solved by technology

However, there are still the following deficiencies: (1) Since the HTTP protocol is stateless, the client's requests for each index file and fragment file of the same HLS content are independent and fragmented, and the cache server cannot complete the request. Track a playback behavior of a client; (2) Since the tail of the HTTP URL is portable? , &, # and other symbols with specific meanings, the current back-to-source method does not define a unified processing rule, which leads to low docking efficiency between the cache node and the source site; Unrelated small files are stored, which will lead to more disk fragments on the cache server, low space utilization, small operation granularity, and low read and write performance. When the file descriptor overhead of the cache server is too large, it will also cause security risks

Images