Data routing control method and device, data analysis processing method and device, equipment and medium

Abstract

The invention discloses a data routing control method and device, a data analysis processing method and device, equipment and a medium, and the data routing control method comprises the steps: receiving a user request, and analyzing an interface address in request data; inquiring a routing table to determine a main route matched with the interface address, forwarding the request data to a first server pointed by a target address of the main route and also to a second server pointed by a bypass address set for the main route, and enabling the first server and the second server to respectively run different version programs of the same online service; monitoring a first version data packet returned after the first server responds, and sending a copy of the first version data packet to a preset test server; and intercepting a second version data packet returned after the second server responds, and only forwarding the second version data packet to the test server. According to the application, the same user request based on the real application demand is asynchronously forwarded to the server where the two version programs of the same online service are located, so that a software comparison test scheme is more effective.

Description

technical field [0001] The embodiments of the present application relate to the field of network data monitoring, in particular to a data routing control method and its device, computer equipment and storage media, and also to a data analysis and processing method. Background technique [0002] The development of Internet online services often requires various effect verifications. There are many effect verification algorithm models in the prior art, which can meet most of the verification requirements. When performing effect verification based on algorithm models, A / B testing is often used to compare and analyze the operation effects of two or more algorithm models, and obtain test result data for comparative analysis. [0003] The implementation of A / B testing means that there are at least two versions of running instances of the same online service, usually distributed in two servers, which is often used when the program of the online service is Test online services to c...

AI Technical Summary

Problems solved by technology

[0006] In theory, the purpose of comparative testing can be achieved by manually providing the same request parameters for different versions of online services. For example, under existing technical conditions, the results of the two models are usually observed by manually sending the same request. Understandably, this method is time-consuming It is laborious, extremely inefficient, and often cannot achieve a more comprehensive test because there are not enough samples. A better way is to try to make different versions of online services respond to real user requests online to assist in the execution of A / B tests.

Images