Chisel: Reshaping Queries to Trim Latency in Key-Value Stores

Robert Birke, Juan F. Perez, Sonia Ben Mokhtar, Navaneeth Rameshan, Lydia Y. Chen

Resultado de la investigación: Tipos de Contribuciónes en ConferenciaPaper

Resumen

It is challenging for key-value data stores to trim user (tail) latency of requests as the workloads are observed to have skewed number of key-value pairs and commonly retrieved via multiget operation, i.e., all keys at the same time. In this paper we present Chisel, a novel client side solution to efficiently reshape the query size at the data store by adaptively splitting big requests into chunks to reap the benefits of parallelism and merge small requests into a single query to amortize latency overheads per request. We derive a novel layered queueing model that can quickly and approximately steer the decisions of Chisel. We extensively evaluate Chisel on memcached clusters hosted on a testbed, across a large number of scenarios with different workloads and system configurations. Our evaluation results show that Chisel can overturn the inherent high variability of requests into a judicious operational region, showcasing significant gains for the mean and 95th percentile of user perceived latency, compared to the state-of-art query processing policy.
Idioma originalEnglish (US)
Número de páginas10
DOI
EstadoPublished - sep 13 2019
Evento2019 IEEE International Conference on Autonomic Computing (ICAC) - Umea
Duración: jun 16 2019jun 20 2019
https://ieeexplore.ieee.org/document/8831208

Conference

Conference2019 IEEE International Conference on Autonomic Computing (ICAC)
Título abreviado2019 IEEE
PaísSwitzerland
CiudadUmea
Período6/16/196/20/19
Dirección de internet

Citar esto

Birke, R., Perez, J. F., Mokhtar, S. B., Rameshan, N., & Chen, L. Y. (2019). Chisel: Reshaping Queries to Trim Latency in Key-Value Stores. Papel presentado en 2019 IEEE International Conference on Autonomic Computing (ICAC), Umea, . https://doi.org/10.1109/icac.2019.00016
Birke, Robert ; Perez, Juan F. ; Mokhtar, Sonia Ben ; Rameshan, Navaneeth ; Chen, Lydia Y. / Chisel: Reshaping Queries to Trim Latency in Key-Value Stores. Papel presentado en 2019 IEEE International Conference on Autonomic Computing (ICAC), Umea, .10 p.
@conference{8c153f7127134249acf58454320c854b,
title = "Chisel: Reshaping Queries to Trim Latency in Key-Value Stores",
abstract = "It is challenging for key-value data stores to trim user (tail) latency of requests as the workloads are observed to have skewed number of key-value pairs and commonly retrieved via multiget operation, i.e., all keys at the same time. In this paper we present Chisel, a novel client side solution to efficiently reshape the query size at the data store by adaptively splitting big requests into chunks to reap the benefits of parallelism and merge small requests into a single query to amortize latency overheads per request. We derive a novel layered queueing model that can quickly and approximately steer the decisions of Chisel. We extensively evaluate Chisel on memcached clusters hosted on a testbed, across a large number of scenarios with different workloads and system configurations. Our evaluation results show that Chisel can overturn the inherent high variability of requests into a judicious operational region, showcasing significant gains for the mean and 95th percentile of user perceived latency, compared to the state-of-art query processing policy.",
author = "Robert Birke and Perez, {Juan F.} and Mokhtar, {Sonia Ben} and Navaneeth Rameshan and Chen, {Lydia Y.}",
year = "2019",
month = "9",
day = "13",
doi = "10.1109/icac.2019.00016",
language = "English (US)",
note = "2019 IEEE International Conference on Autonomic Computing (ICAC), 2019 IEEE ; Conference date: 16-06-2019 Through 20-06-2019",
url = "https://ieeexplore.ieee.org/document/8831208",

}

Birke, R, Perez, JF, Mokhtar, SB, Rameshan, N & Chen, LY 2019, 'Chisel: Reshaping Queries to Trim Latency in Key-Value Stores', Papel presentado en 2019 IEEE International Conference on Autonomic Computing (ICAC), Umea, 6/16/19 - 6/20/19. https://doi.org/10.1109/icac.2019.00016

Chisel: Reshaping Queries to Trim Latency in Key-Value Stores. / Birke, Robert; Perez, Juan F.; Mokhtar, Sonia Ben; Rameshan, Navaneeth; Chen, Lydia Y.

2019. Papel presentado en 2019 IEEE International Conference on Autonomic Computing (ICAC), Umea, .

Resultado de la investigación: Tipos de Contribuciónes en ConferenciaPaper

TY - CONF

T1 - Chisel: Reshaping Queries to Trim Latency in Key-Value Stores

AU - Birke, Robert

AU - Perez, Juan F.

AU - Mokhtar, Sonia Ben

AU - Rameshan, Navaneeth

AU - Chen, Lydia Y.

PY - 2019/9/13

Y1 - 2019/9/13

N2 - It is challenging for key-value data stores to trim user (tail) latency of requests as the workloads are observed to have skewed number of key-value pairs and commonly retrieved via multiget operation, i.e., all keys at the same time. In this paper we present Chisel, a novel client side solution to efficiently reshape the query size at the data store by adaptively splitting big requests into chunks to reap the benefits of parallelism and merge small requests into a single query to amortize latency overheads per request. We derive a novel layered queueing model that can quickly and approximately steer the decisions of Chisel. We extensively evaluate Chisel on memcached clusters hosted on a testbed, across a large number of scenarios with different workloads and system configurations. Our evaluation results show that Chisel can overturn the inherent high variability of requests into a judicious operational region, showcasing significant gains for the mean and 95th percentile of user perceived latency, compared to the state-of-art query processing policy.

AB - It is challenging for key-value data stores to trim user (tail) latency of requests as the workloads are observed to have skewed number of key-value pairs and commonly retrieved via multiget operation, i.e., all keys at the same time. In this paper we present Chisel, a novel client side solution to efficiently reshape the query size at the data store by adaptively splitting big requests into chunks to reap the benefits of parallelism and merge small requests into a single query to amortize latency overheads per request. We derive a novel layered queueing model that can quickly and approximately steer the decisions of Chisel. We extensively evaluate Chisel on memcached clusters hosted on a testbed, across a large number of scenarios with different workloads and system configurations. Our evaluation results show that Chisel can overturn the inherent high variability of requests into a judicious operational region, showcasing significant gains for the mean and 95th percentile of user perceived latency, compared to the state-of-art query processing policy.

UR - http://www.mendeley.com/research/chisel-reshaping-queries-trim-latency-keyvalue-stores

UR - http://www.mendeley.com/research/chisel-reshaping-queries-trim-latency-keyvalue-stores

U2 - 10.1109/icac.2019.00016

DO - 10.1109/icac.2019.00016

M3 - Paper

ER -

Birke R, Perez JF, Mokhtar SB, Rameshan N, Chen LY. Chisel: Reshaping Queries to Trim Latency in Key-Value Stores. 2019. Papel presentado en 2019 IEEE International Conference on Autonomic Computing (ICAC), Umea, . https://doi.org/10.1109/icac.2019.00016