Improving LSM-trie performance by parallel search
- Publikationstyp:
- Zeitschriftenaufsatz
- Metadaten:
-
- Autoren
- Wen Cheng
- Tao Guo
- Lingfang Zeng
- Yang Wang
- Lars Nagel
- Tim Suess
- Andre Brinkmann
- Autoren-URL
- https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=fis-test-1&SrcAuth=WosAPI&KeyUT=WOS:000554912900001&DestLinkType=FullRecord&DestApp=WOS_CPL
- DOI
- 10.1002/spe.2875
- eISSN
- 1097-024X
- Externe Identifier
- Clarivate Analytics Document Solution ID: NJ8HF
- ISSN
- 0038-0644
- Ausgabe der Veröffentlichung
- 10
- Zeitschrift
- SOFTWARE-PRACTICE & EXPERIENCE
- Schlüsselwörter
- key-value storage
- LSM-tree
- parallel search
- Paginierung
- 1952 - 1965
- Datum der Veröffentlichung
- 2020
- Status
- Published
- Titel
- Improving LSM-trie performance by parallel search
- Sub types
- Article
- Ausgabe der Zeitschrift
- 50
Datenquelle: Web of Science (Lite)
- Andere Metadatenquellen:
-
- Abstract
- <jats:title>Summary</jats:title><jats:p>LSM‐trie‐based key‐value (KV) store is often used to manage an ultralarge dataset in reality by introducing a number of sublevels at each level, its linear growth pattern can fairly reduce the write amplification in store operations. Although this design is effective for the write operation, the last level holds a large proportion of KV items, leading to the extreme imbalance of data distribution. Therefore, to support efficient read, we need to carefully consider this imbalance. On the other hand, to ensure that acquired data is latest, the LSM‐trie needs to search the dataset at different levels one by one, and this search method may take a lot of unnecessary time. When the number of items is ultralarge, the random lookup performance may be poor due to the imbalance data distribution. To address this issue, we improve the read performance of the LSM‐trie by changing its <jats:italic>serial search</jats:italic> to <jats:italic>parallel search</jats:italic>, using two threads to simultaneously search at the last level and other levels, respectively. Our experiment results show that the read performance of the LSM‐trie can be improved up to <jats:styled-content>98.35<jats:italic>%</jats:italic></jats:styled-content> and on average <jats:styled-content>71.55<jats:italic>%</jats:italic></jats:styled-content>.</jats:p>
- Autoren
- Wen Cheng
- Tao Guo
- Lingfang Zeng
- Yang Wang
- Lars Nagel
- Tim Süß
- André Brinkmann
- DOI
- 10.1002/spe.2875
- eISSN
- 1097-024X
- ISSN
- 0038-0644
- Ausgabe der Veröffentlichung
- 10
- Zeitschrift
- Software: Practice and Experience
- Sprache
- en
- Online publication date
- 2020
- Paginierung
- 1952 - 1965
- Datum der Veröffentlichung
- 2020
- Status
- Published
- Herausgeber
- Wiley
- Herausgeber URL
- http://dx.doi.org/10.1002/spe.2875
- Datum der Datenerfassung
- 2023
- Titel
- Improving LSM‐trie performance by parallel search
- Ausgabe der Zeitschrift
- 50
Datenquelle: Crossref
- Autoren
- Wen Cheng
- Tao Guo
- Lingfang Zeng
- Yang Wang
- Lars Nagel
- Tim Süß
- André Brinkmann
- Zeitschrift
- Softw. Pract. Exp.
- Artikelnummer
- 10
- Paginierung
- 1952 - 1965
- Datum der Veröffentlichung
- 2020
- Titel
- Improving LSM-trie performance by parallel search.
- Ausgabe der Zeitschrift
- 50
Datenquelle: DBLP
- Autoren
- Wen Cheng
- Tao Guo
- Lingfang Zeng
- Yang Wang
- Lars Nagel
- Tim Süß
- André Brinkmann
- DOI
- 10.1002/spe.2875
- Zeitschrift
- Software: Practice and Experience
- Artikelnummer
- 10
- Paginierung
- 1952 - 1965
- Datum der Veröffentlichung
- 2020
- Datum der Datenerfassung
- 2020
- Titel
- Improving LSM-trie performance by parallel search
- Ausgabe der Zeitschrift
- 50
Datenquelle: Manual
- Beziehungen:
- Eigentum von