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
Os tempos de resposta das unidades de estado sólido (SSDs) diminuíram drasticamente devido ao uso crescente de dispositivos expressos de memória não volátil (NVMe). Esses dispositivos têm tempos de resposta inferiores a 100 microssegundos, em média. Os tempos de resposta dos sistemas all-flash array também diminuíram drasticamente com o uso de SSDs NVMe. No entanto, existem aplicações, especialmente infraestruturas de desktops virtuais e sistemas de bancos de dados na memória, que exigem sistemas de armazenamento com tempos de resposta ainda mais curtos. Suas cargas de trabalho tendem a conter muitas concentrações de entrada-saída (IO), que são agregações de acessos de IO. Eles têm como alvo regiões estreitas do volume de armazenamento e podem continuar por até uma hora. Essas regiões estreitas ocupam uma pequena porcentagem da capacidade numérica da unidade lógica, são o alvo da maioria dos acessos de E/S e aparecem em endereços de blocos lógicos imprevisíveis. Para reduzir drasticamente os tempos de resposta para essas cargas de trabalho, desenvolvemos um sistema automatizado de armazenamento em camadas chamado “armazenamento automatizado em camadas com memória rápida e armazenamento flash lento” (ATSMF), no qual os dados nas regiões alvo são migrados entre dispositivos de armazenamento, dependendo do tempo restante previsto. duração da concentração. O ambiente assumido é um servidor com memória não volátil e SSDs conectados diretamente, com as aplicações do usuário executadas no servidor, pois isso reduz o tempo médio de resposta. Nosso sistema prevê o efeito da migração usando os valores monitorados anteriormente do aumento no tempo de resposta durante a migração e a mudança no tempo de resposta após a migração. Esses valores serão consistentes para cada tipo de carga de trabalho se o sistema for criado usando memória não volátil e SSDs. Em particular, o sistema prevê a duração restante de uma concentração de IO, calcula o aumento esperado do tempo de resposta durante a migração e a diminuição esperada do tempo de resposta após a migração, e migra os dados nas regiões alvo se a soma da diminuição do tempo de resposta após a migração excede a soma do aumento do tempo de resposta durante a migração. Resultados experimentais indicam que o ATSMF é pelo menos 20% mais rápido que apenas o armazenamento flash e que sua taxa de acesso à memória é superior a 50%.
Kazuichi OE
FUJITSU LABORATORIES LTD.
Mitsuru SATO
FUJITSU LABORATORIES LTD.
Takeshi NANRI
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.
Copiar
Kazuichi OE, Mitsuru SATO, Takeshi NANRI, "ATSMF: Automated Tiered Storage with Fast Memory and Slow Flash Storage to Improve Response Time with Concentrated Input-Output (IO) Workloads" in IEICE TRANSACTIONS on Information,
vol. E101-D, no. 12, pp. 2889-2901, December 2018, doi: 10.1587/transinf.2018PAP0005.
Abstract: The response times of solid state drives (SSDs) have decreased dramatically due to the growing use of non-volatile memory express (NVMe) devices. Such devices have response times of less than 100 micro seconds on average. The response times of all-flash-array systems have also decreased dramatically through the use of NVMe SSDs. However, there are applications, particularly virtual desktop infrastructure and in-memory database systems, that require storage systems with even shorter response times. Their workloads tend to contain many input-output (IO) concentrations, which are aggregations of IO accesses. They target narrow regions of the storage volume and can continue for up to an hour. These narrow regions occupy a few percent of the logical unit number capacity, are the target of most IO accesses, and appear at unpredictable logical block addresses. To drastically reduce the response times for such workloads, we developed an automated tiered storage system called “automated tiered storage with fast memory and slow flash storage” (ATSMF) in which the data in targeted regions are migrated between storage devices depending on the predicted remaining duration of the concentration. The assumed environment is a server with non-volatile memory and directly attached SSDs, with the user applications executed on the server as this reduces the average response time. Our system predicts the effect of migration by using the previously monitored values of the increase in response time during migration and the change in response time after migration. These values are consistent for each type of workload if the system is built using both non-volatile memory and SSDs. In particular, the system predicts the remaining duration of an IO concentration, calculates the expected response-time increase during migration and the expected response-time decrease after migration, and migrates the data in the targeted regions if the sum of response-time decrease after migration exceeds the sum of response-time increase during migration. Experimental results indicate that ATSMF is at least 20% faster than flash storage only and that its memory access ratio is more than 50%.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018PAP0005/_p
Copiar
@ARTICLE{e101-d_12_2889,
author={Kazuichi OE, Mitsuru SATO, Takeshi NANRI, },
journal={IEICE TRANSACTIONS on Information},
title={ATSMF: Automated Tiered Storage with Fast Memory and Slow Flash Storage to Improve Response Time with Concentrated Input-Output (IO) Workloads},
year={2018},
volume={E101-D},
number={12},
pages={2889-2901},
abstract={The response times of solid state drives (SSDs) have decreased dramatically due to the growing use of non-volatile memory express (NVMe) devices. Such devices have response times of less than 100 micro seconds on average. The response times of all-flash-array systems have also decreased dramatically through the use of NVMe SSDs. However, there are applications, particularly virtual desktop infrastructure and in-memory database systems, that require storage systems with even shorter response times. Their workloads tend to contain many input-output (IO) concentrations, which are aggregations of IO accesses. They target narrow regions of the storage volume and can continue for up to an hour. These narrow regions occupy a few percent of the logical unit number capacity, are the target of most IO accesses, and appear at unpredictable logical block addresses. To drastically reduce the response times for such workloads, we developed an automated tiered storage system called “automated tiered storage with fast memory and slow flash storage” (ATSMF) in which the data in targeted regions are migrated between storage devices depending on the predicted remaining duration of the concentration. The assumed environment is a server with non-volatile memory and directly attached SSDs, with the user applications executed on the server as this reduces the average response time. Our system predicts the effect of migration by using the previously monitored values of the increase in response time during migration and the change in response time after migration. These values are consistent for each type of workload if the system is built using both non-volatile memory and SSDs. In particular, the system predicts the remaining duration of an IO concentration, calculates the expected response-time increase during migration and the expected response-time decrease after migration, and migrates the data in the targeted regions if the sum of response-time decrease after migration exceeds the sum of response-time increase during migration. Experimental results indicate that ATSMF is at least 20% faster than flash storage only and that its memory access ratio is more than 50%.},
keywords={},
doi={10.1587/transinf.2018PAP0005},
ISSN={1745-1361},
month={December},}
Copiar
TY - JOUR
TI - ATSMF: Automated Tiered Storage with Fast Memory and Slow Flash Storage to Improve Response Time with Concentrated Input-Output (IO) Workloads
T2 - IEICE TRANSACTIONS on Information
SP - 2889
EP - 2901
AU - Kazuichi OE
AU - Mitsuru SATO
AU - Takeshi NANRI
PY - 2018
DO - 10.1587/transinf.2018PAP0005
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E101-D
IS - 12
JA - IEICE TRANSACTIONS on Information
Y1 - December 2018
AB - The response times of solid state drives (SSDs) have decreased dramatically due to the growing use of non-volatile memory express (NVMe) devices. Such devices have response times of less than 100 micro seconds on average. The response times of all-flash-array systems have also decreased dramatically through the use of NVMe SSDs. However, there are applications, particularly virtual desktop infrastructure and in-memory database systems, that require storage systems with even shorter response times. Their workloads tend to contain many input-output (IO) concentrations, which are aggregations of IO accesses. They target narrow regions of the storage volume and can continue for up to an hour. These narrow regions occupy a few percent of the logical unit number capacity, are the target of most IO accesses, and appear at unpredictable logical block addresses. To drastically reduce the response times for such workloads, we developed an automated tiered storage system called “automated tiered storage with fast memory and slow flash storage” (ATSMF) in which the data in targeted regions are migrated between storage devices depending on the predicted remaining duration of the concentration. The assumed environment is a server with non-volatile memory and directly attached SSDs, with the user applications executed on the server as this reduces the average response time. Our system predicts the effect of migration by using the previously monitored values of the increase in response time during migration and the change in response time after migration. These values are consistent for each type of workload if the system is built using both non-volatile memory and SSDs. In particular, the system predicts the remaining duration of an IO concentration, calculates the expected response-time increase during migration and the expected response-time decrease after migration, and migrates the data in the targeted regions if the sum of response-time decrease after migration exceeds the sum of response-time increase during migration. Experimental results indicate that ATSMF is at least 20% faster than flash storage only and that its memory access ratio is more than 50%.
ER -