公开(公告)号：US20220414075A1

公开(公告)日：2022-12-29

申请号：US17897008

申请日：2022-08-26

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Guoliang LI ,  Tianqing WANG ,  Shifu LI ,  Lingtao LI ,  Kun LI

IPC: G06F16/21 ,  G06F16/242 ,  G06F16/2453

Abstract: Disclosed is a method for estimating database management system performance, in which a performance change ratio of a DBMS can be determined once a first knob group, a second knob group, and a data volume of active data in data managed by the DBMS are obtained, without actually configuring the second knob group in the DBMS, executing a job by the DBMS, and then observing the execution. In other words, the performance change ratio of the DBMS can be estimated without interacting with the DBMS. DBMS security can be ensured, performance measurement approaches are provided for self-tuning and self-management of the DBMS, and reliable and stable running of the DBMS is ensured.

公开(公告)号：US20240152807A1

公开(公告)日：2024-05-09

申请号：US18398359

申请日：2023-12-28

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Shifu LI ,  Tianqing WANG ,  Wen NIE

IPC: G06N20/00 ,  G06F16/2455

CPC classification number: G06N20/00 ,  G06F16/24564

Abstract: A data processing method applied to a database system is disclosed. The method includes: obtaining a model training request, where the model training request includes a plurality of training samples and a model training policy, and the plurality of training samples is grouped into N training sample groups; generating an execution plan of the model training policy and an estimated execution cost of the execution plan executed by the database system; obtaining, based on the estimated execution cost, M training sample groups in the N training sample groups; training a to-be-trained model in parallel by using the M training sample groups, to obtain M pieces of parameter update data; and updating the to-be-trained model based on the M pieces of parameter update data, to obtain a trained model. This method reduces time overheads of the model training.