2024-03-29T22:40:38Zhttps://ipsj.ixsq.nii.ac.jp/ej/?action=repository_oaipmhoai:ipsj.ixsq.nii.ac.jp:001788702024-03-29T05:26:34Z01164:01867:09088:09164
Operating System Integrated Tracer for Database Management System in Big Data ProcessingOperating System Integrated Tracer for Database Management System in Big Data Processingengビッグデータhttp://id.nii.ac.jp/1001/00178782/Technical Reporthttps://ipsj.ixsq.nii.ac.jp/ej/?action=repository_action_common_download&item_id=178870&item_no=1&attribute_id=1&file_no=1Copyright (c) 2017 by the Information Processing Society of JapanResearch & Development Group, Hitachi, Ltd.Research & Development Group, Hitachi, Ltd.Research & Development Group, Hitachi, Ltd.Keiko, TanigawaKazuya, HisakiKeisuke, HatasakiBig data processing has become more important recently. Various data I/O processing acceleration technologies have been developed for it, for example, many-core CPUs, large capacity memory, flash storage, and parallel processing. However, the current OS is designed to run various applications on commodity hardware, and the OS can be a bottleneck in big data processing. Therefore, in big data processing on high performance hardware, an SQL bottleneck analysis needs to drill down into the OS layer to confirm which is the cause of bottleneck, the SQL or the OS. This paper proposes an OS integrated trace tool which analyzes OS functions related to the behavior of SQL data processing. The results of the experiments show that it can extract the kernel functions from SQL and the OS trace within a practical time.Big data processing has become more important recently. Various data I/O processing acceleration technologies have been developed for it, for example, many-core CPUs, large capacity memory, flash storage, and parallel processing. However, the current OS is designed to run various applications on commodity hardware, and the OS can be a bottleneck in big data processing. Therefore, in big data processing on high performance hardware, an SQL bottleneck analysis needs to drill down into the OS layer to confirm which is the cause of bottleneck, the SQL or the OS. This paper proposes an OS integrated trace tool which analyzes OS functions related to the behavior of SQL data processing. The results of the experiments show that it can extract the kernel functions from SQL and the OS trace within a practical time.AN10444176研究報告システムソフトウェアとオペレーティング・システム(OS)2017-OS-14012152017-05-092188-87952017-04-28