The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Dans des études précédentes, nous avons déterminé que les charges de travail contiennent souvent de nombreuses concentrations d’entrées-sorties (IO). De telles concentrations sont des agrégations d’accès IO. Ils apparaissent dans des régions étroites d’un volume de stockage et durent jusqu’à environ une heure. Ces régions étroites occupent un petit pourcentage de la capacité du nombre d'unités logiques, incluent la plupart des accès IO et apparaissent à des adresses de blocs logiques imprévisibles. Nous avons étudié ces charges de travail en nous concentrant sur la régularité au niveau des pages et avons constaté qu'elles incluent souvent peu de régularités. Cela signifie qu'une simple mise en cache peut ne pas réduire suffisamment le temps de réponse de ces charges de travail, car l'algorithme de migration du cache utilise une régularité au niveau de la page. Nous avons précédemment développé un système de hiérarchisation automatisée du stockage à la volée (OTF-AST) composé d'un SSD et d'un disque dur. L'algorithme de migration identifie les concentrations d'E/S de durée moyennement longue et les migre du disque dur vers le SSD. Cela signifie qu’il y a peu ou pas de réduction du temps de réponse lorsque la charge de travail comprend peu de telles concentrations. Nous avons maintenant développé un système de stockage hybride composé d'un lecteur de cache avec un SSD et un disque dur et d'un SSD multiniveau utilisant OTF-AST, appelé « OTF-AST avec mise en cache ». Le schéma OTF-AST gère les accès IO qui produisent des concentrations d’IO de durée moyennement longue, tandis que le schéma de mise en cache gère les accès IO restants. Des expériences ont montré que le temps de réponse moyen de notre système était 45 % supérieur à celui de Facebook FlashCache sur une charge de travail Microsoft Research Cambridge.
Kazuichi OE
FUJITSU LABORATORIES LTD
Takeshi NANRI
Kyushu University
Koji OKAMURA
Kyushu University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Kazuichi OE, Takeshi NANRI, Koji OKAMURA, "Hybrid Storage System Consisting of Cache Drive and Multi-Tier SSD for Improved IO Access when IO is Concentrated" in IEICE TRANSACTIONS on Information,
vol. E102-D, no. 9, pp. 1715-1730, September 2019, doi: 10.1587/transinf.2018EDP7253.
Abstract: In previous studies, we determined that workloads often contain many input-output (IO) concentrations. Such concentrations are aggregations of IO accesses. They appear in narrow regions of a storage volume and continue for durations of up to about an hour. These narrow regions occupy a small percentage of the logical unit number capacity, include most IO accesses, and appear at unpredictable logical block addresses. We investigated these workloads by focusing on page-level regularity and found that they often include few regularities. This means that simple caching may not reduce the response time for these workloads sufficiently because the cache migration algorithm uses page-level regularity. We previously developed an on-the-fly automated storage tiering (OTF-AST) system consisting of an SSD and an HDD. The migration algorithm identifies IO concentrations with moderately long durations and migrates them from the HDD to the SSD. This means that there is little or no reduction in the response time when the workload includes few such concentrations. We have now developed a hybrid storage system consisting of a cache drive with an SSD and HDD and a multi-tier SSD that uses OTF-AST, called “OTF-AST with caching.” The OTF-AST scheme handles the IO accesses that produce moderately long duration IO concentrations while the caching scheme handles the remaining IO accesses. Experiments showed that the average response time for our system was 45% that of Facebook FlashCache on a Microsoft Research Cambridge workload.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7253/_p
Copier
@ARTICLE{e102-d_9_1715,
author={Kazuichi OE, Takeshi NANRI, Koji OKAMURA, },
journal={IEICE TRANSACTIONS on Information},
title={Hybrid Storage System Consisting of Cache Drive and Multi-Tier SSD for Improved IO Access when IO is Concentrated},
year={2019},
volume={E102-D},
number={9},
pages={1715-1730},
abstract={In previous studies, we determined that workloads often contain many input-output (IO) concentrations. Such concentrations are aggregations of IO accesses. They appear in narrow regions of a storage volume and continue for durations of up to about an hour. These narrow regions occupy a small percentage of the logical unit number capacity, include most IO accesses, and appear at unpredictable logical block addresses. We investigated these workloads by focusing on page-level regularity and found that they often include few regularities. This means that simple caching may not reduce the response time for these workloads sufficiently because the cache migration algorithm uses page-level regularity. We previously developed an on-the-fly automated storage tiering (OTF-AST) system consisting of an SSD and an HDD. The migration algorithm identifies IO concentrations with moderately long durations and migrates them from the HDD to the SSD. This means that there is little or no reduction in the response time when the workload includes few such concentrations. We have now developed a hybrid storage system consisting of a cache drive with an SSD and HDD and a multi-tier SSD that uses OTF-AST, called “OTF-AST with caching.” The OTF-AST scheme handles the IO accesses that produce moderately long duration IO concentrations while the caching scheme handles the remaining IO accesses. Experiments showed that the average response time for our system was 45% that of Facebook FlashCache on a Microsoft Research Cambridge workload.},
keywords={},
doi={10.1587/transinf.2018EDP7253},
ISSN={1745-1361},
month={September},}
Copier
TY - JOUR
TI - Hybrid Storage System Consisting of Cache Drive and Multi-Tier SSD for Improved IO Access when IO is Concentrated
T2 - IEICE TRANSACTIONS on Information
SP - 1715
EP - 1730
AU - Kazuichi OE
AU - Takeshi NANRI
AU - Koji OKAMURA
PY - 2019
DO - 10.1587/transinf.2018EDP7253
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2019
AB - In previous studies, we determined that workloads often contain many input-output (IO) concentrations. Such concentrations are aggregations of IO accesses. They appear in narrow regions of a storage volume and continue for durations of up to about an hour. These narrow regions occupy a small percentage of the logical unit number capacity, include most IO accesses, and appear at unpredictable logical block addresses. We investigated these workloads by focusing on page-level regularity and found that they often include few regularities. This means that simple caching may not reduce the response time for these workloads sufficiently because the cache migration algorithm uses page-level regularity. We previously developed an on-the-fly automated storage tiering (OTF-AST) system consisting of an SSD and an HDD. The migration algorithm identifies IO concentrations with moderately long durations and migrates them from the HDD to the SSD. This means that there is little or no reduction in the response time when the workload includes few such concentrations. We have now developed a hybrid storage system consisting of a cache drive with an SSD and HDD and a multi-tier SSD that uses OTF-AST, called “OTF-AST with caching.” The OTF-AST scheme handles the IO accesses that produce moderately long duration IO concentrations while the caching scheme handles the remaining IO accesses. Experiments showed that the average response time for our system was 45% that of Facebook FlashCache on a Microsoft Research Cambridge workload.
ER -