公开(公告)号：US20210297305A1

公开(公告)日：2021-09-23

申请号：US17342659

申请日：2021-06-09

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Xiaoyun SI ,  Dewei BAO ,  Liang ZHANG ,  Zhenhang SUN

IPC: H04L12/24

Abstract: A fault root cause identification method, apparatus, and device. For a failure flow that occurs when a connectivity fault for access in a network occurs, a target success flow that has a high similarity with the failure flow is determined from a plurality of success flows in the network based on the failure flow. Then, a target fault root cause of the failure flow is obtained based on the failure flow, the target success flow, and a trained first machine learning model. In this way, the target success flow related to the failure flow is determined from the plurality of success flows in the network, and the first machine learning model trained by using a large quantity of success flows and failure flows whose feature indicators are slightly different from each other is used.

公开(公告)号：US20180367367A1

公开(公告)日：2018-12-20

申请号：US16111234

申请日：2018-08-24

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Dewei BAO ,  Donghong PENG ,  Yaodong GUAN ,  Shihao LI ,  Zhiqiang LI

IPC: H04L12/24

Abstract: A network log time alignment method and an apparatus are provided. The method includes: obtaining log data stored in multiple network devices; parsing the log data to obtain features of each network device; determining associated features of the multiple network devices according to the features of each network device, where the associated features are features that are of the network devices and that have an association in space; and then, performing alignment on log time in multiple pieces of log data according to the associated features. Therefore, associated features of related devices of a network fault can be found, and log time are automatically aligned according to the associated features, thereby improving the efficiency and the accuracy of the fault location and sectionalization.