{"updated":"2025-01-19T18:04:47.193504+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00210346","sets":["581:10433:10436"]},"path":["10436"],"owner":"44499","recid":"210346","title":["HeteroTSDB：異種分散KVS間の自動階層化による高性能な時系列データベース"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-03-15"},"_buckets":{"deposit":"361446a9-2f41-4338-ba0f-11b08a2d95ef"},"_deposit":{"id":"210346","pid":{"type":"depid","value":"210346","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"HeteroTSDB：異種分散KVS間の自動階層化による高性能な時系列データベース","author_link":["532299","532307","532304","532302","532298","532308","532297","532309","532301","532305","532296","532303","532306","532300"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"HeteroTSDB：異種分散KVS間の自動階層化による高性能な時系列データベース"},{"subitem_title":"HeteroTSDB: A High Performance Time Series Database for Automated Data Tiering in Heterogeneous Distributed KVSs","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"[特集：快適な運用管理を支えるインターネットと運用技術] key value store，monitoring，time series database，redis，cassandra","subitem_subject_scheme":"Other"}]},"item_type_id":"2","publish_date":"2021-03-15","item_2_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"さくらインターネット株式会社さくらインターネット研究所／京都大学情報学研究科"},{"subitem_text_value":"株式会社はてな"},{"subitem_text_value":"株式会社はてな"},{"subitem_text_value":"Nature株式会社"},{"subitem_text_value":"株式会社野村総合研究所"},{"subitem_text_value":"トヨタ自動車株式会社"},{"subitem_text_value":"さくらインターネット株式会社さくらインターネット研究所"}]},"item_2_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"SAKURA internet Research Center, SAKURA internet Inc. / Graduate School of Infomatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Hatena Co., Ltd.","subitem_text_language":"en"},{"subitem_text_value":"Hatena Co., Ltd.","subitem_text_language":"en"},{"subitem_text_value":"Nature Inc.","subitem_text_language":"en"},{"subitem_text_value":"Nomura Research Institute, Ltd.","subitem_text_language":"en"},{"subitem_text_value":"Toyota Motor Corporation","subitem_text_language":"en"},{"subitem_text_value":"SAKURA internet Research Center, SAKURA internet Inc.","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/210346/files/IPSJ-JNL6203003.pdf","label":"IPSJ-JNL6203003.pdf"},"date":[{"dateType":"Available","dateValue":"2023-03-15"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-JNL6203003.pdf","filesize":[{"value":"933.2 kB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"8"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"35c6169f-1ce0-4b99-9a25-61783bd40465","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2021 by the Information Processing Society of Japan"}]},"item_2_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"坪内, 佑樹"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"脇坂, 朝人"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"濱田, 健"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"松木, 雅幸"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"小林, 隆浩"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"阿部, 博"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"松本, 亮介"}],"nameIdentifiers":[{}]}]},"item_2_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Yuuki, Tsubouchi","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Asato, Wakisaka","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Ken, Hamada","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Masayuki, Matsuki","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Takahiro, Kobayashi","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Hiroshi, Abe","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Ryosuke, Matsumoto","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_2_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN00116647","subitem_source_identifier_type":"NCID"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_6501","resourcetype":"journal article"}]},"item_2_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"1882-7764","subitem_source_identifier_type":"ISSN"}]},"item_2_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"クラウド上のシステムの大規模化にともない，CPU利用率などのシステムの各構成要素の状態を把握するために，大量の時系列データを保存する必要がある．そのために，時系列データを保存するデータベースにはデータの挿入処理とデータ保存の効率化と挿入スケールアウト性の向上が求められる．既存技術は，挿入スケールアウト性を高めるために広く利用されているディスクベースの分散KVS（Key Value Store）を利用する．しかし，ランダムI/Oが低速なディスクへ書き込むという前提があることから，メモリ上でキーを整列させながら挿入可能な平衡木が利用されるが，キーの挿入時に系列数に対して対数時間を要する．すべてのデータをメモリ上に保持するメモリベースKVSであれば，ハッシュ表に基づくデータ構造の利用により定数時間の挿入が可能となる．しかし，メモリは容量単価が大きいことから，データを長期間保存するには不向きである．本論文では，メモリベースKVSとディスクベースKVSを階層化する高性能な時系列データベースHeteroTSDBを提案する．HeteroTSDBは，メモリベースKVS上に系列名をキーとして，系列本体をバリューとしたハッシュ表を構成することにより，系列数に対して定数時間でデータを挿入する．加えて，系列を格納するキーにTTL（Time To Live）によるタイマを設定し，古くなったデータを系列単位でまとめてディスクベースKVSへ移動させることにより，データ保存のための容量単価を低減させている．実験の結果，ディスクベースKVSを利用した既存の時系列データベースであるKairosDBと比較し，HeteroTSDBは3.98倍の挿入スループット向上を達成した．","subitem_description_type":"Other"}]},"item_2_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"As the scale of systems on the cloud grows, there is a need to store a significant amount of time series data to understand the status of each component of systems, such as CPU utilization. Time series databases require high insertion scale-out property and high efficient utilization of hardware resources for insertion and storage. Existing technologies utilize disk-based distributed KVSs (Key Value Stores), which is widely used to achieve insertion scale-put property. Since random I/Os to disk are slower than to memory, the disk-based KVSs adopt an equilibrium tree, that can be inserted while aligning the keys in memory. However, it takes logarithmic time for the number of time series. Memory-based KVSs, where all the data are stored in memory, adopts data structure based on a hash table, which is possible to insert an key in constant time. However, it is not suitable for long-term storage because the unit cost of memory is higher than disk. In this paper, we propose HeteroTSDB, a high performance time series database that tiers a memory-based KVS and a disk-based KVS. HeteroTSDB accelerates the inserting process by constructing a hash table with a series name as a key and series data as a value on the memory-based KVS. Automatic tiering reduces the unit cost of storage capacity by moving the old data from a memory-based KVS to a disk-based KVS. Benchmark experimentation demonstrates that, compared with the existing time series database on a disk-based KVS KairosDB, HeteroTSDB achieves performance improvement of up to 3.98 times in insert throughput.","subitem_description_type":"Other"}]},"item_2_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"828","bibliographic_titles":[{"bibliographic_title":"情報処理学会論文誌"}],"bibliographicPageStart":"818","bibliographicIssueDates":{"bibliographicIssueDate":"2021-03-15","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"3","bibliographicVolumeNumber":"62"}]},"relation_version_is_last":true,"item_2_identifier_registration":{"attribute_name":"ID登録","attribute_value_mlt":[{"subitem_identifier_reg_text":"10.20729/00210244","subitem_identifier_reg_type":"JaLC"}]},"weko_creator_id":"44499"},"created":"2025-01-19T01:11:34.515148+00:00","id":210346,"links":{}}