Method and device for returning data to source

Abstract

Embodiments of the present invention provide a method and device for data back-to-source. The method is applied to an edge node, and the edge node is used for data back-to-source from a core node. The core node stores data synchronized from a global storage node. The method includes: receiving a first data acquisition request on the edge node side, and if the edge node does not have a target file corresponding to the first data acquisition request, then determining the attribute information of the target file; based on the attribute information , logically divide the target file into multiple data blocks; determine one or more target data blocks from the multiple data blocks; determine the core node corresponding to the one or more target data blocks; or core nodes corresponding to multiple target data blocks request target video data. The invention can improve the success rate and efficiency of back-to-source, and save bandwidth and storage resources.

Description

technical field [0001] The present invention relates to the technical field of data processing, and in particular, to a method for returning data to a source and a device for returning data to the source. Background technique [0002] In recent years, with the rapid development of Internet technology, the demand for users to visit major websites is increasing day by day. In order to improve the access response speed, enable users to obtain the requested content faster, and solve the state of network congestion, CDN (Content delivery network, content delivery network) plays an important role. [0003] Back-to-source is an important technology in CDN. Back-to-source means that when the IDC (Internet Data Center, Internet Data Center) node in the CDN that receives the user data request does not have the data requested by the user, the IDC node sends back the data requested by the user. The source node requests the data requested by the user. [0004] In the existing back-to-s...

AI Technical Summary

Problems solved by technology

[0004] In the existing back-to-source mode, when a back-to-source node goes to another back-to-source node for data back-to-source, if the back-to-source node does not have the data that needs to go back to the source, the back-to-source node needs to send data to other back-to-source nodes For data requests, the success rate of returning to the source is low, which greatly wastes bandwidth and storage resources;

[0005] Secondly, in the existing back-to-source method, the server first obtains a list of available back-to-source IP addresses through a certain method (static configuration, dynamic configuration, real-time acquisition and other existing methods); Algorithms (existing methods such as polling and consistent hashing, hereinafter referred to as positioning algorithms) select an IP address to download from this list, and only select an IP address to download the entire file, resulting in subsequent requests for the same file , when requesting undownloaded data, the waiting time is longer, or even freezes occur

Images