Design of DRAM-NAND flash hybrid main memory and Q-learning-based prefetching method

Su Kyung Yoon, Young Sun Youn, Jeong Geun Kim, Shin Dug Kim

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Owing to the increased need for machine learning and artificial intelligence in current cloud computing systems, the amount of data that needs to be processed has exponentially increased. Thus, it is important to optimize memory and storage systems to reduce the energy consumption and execution time of applications. This paper proposes a new Q-learning-based prefetching algorithm for DRAM–NAND flash hybrid main memory architecture. To minimize the computational overheads of learning-based schemes, we have designed two learning policies, namely aggressive learning and lazy learning. The proposed system reduces the energy consumption by about 80% of the memory and storage for Redis, OpenStack Swift which is a cloud computing open source framework and Apache Storm workloads. Further, the overall execution time of workloads in cloud computing applications is reduced by almost half. Using a path generator with a Q-learning-based prefetching algorithm, we realize an increased hit rate of about 21% compared to that with a no-prefetching system, compared to non-prefetching system.

Original languageEnglish
Pages (from-to)5293-5313
Number of pages21
JournalJournal of Supercomputing
Volume74
Issue number10
DOIs
Publication statusPublished - 2018 Oct 1

Fingerprint

NAND
Q-learning
Prefetching
Dynamic random access storage
Flash
Cloud computing
Cloud Computing
Data storage equipment
Energy utilization
Execution Time
Energy Consumption
Workload
Memory architecture
Artificial intelligence
Learning systems
Storage System
Hits
Open Source
Artificial Intelligence
Machine Learning

All Science Journal Classification (ASJC) codes

  • Software
  • Theoretical Computer Science
  • Information Systems
  • Hardware and Architecture

Cite this

Yoon, Su Kyung ; Youn, Young Sun ; Kim, Jeong Geun ; Kim, Shin Dug. / Design of DRAM-NAND flash hybrid main memory and Q-learning-based prefetching method. In: Journal of Supercomputing. 2018 ; Vol. 74, No. 10. pp. 5293-5313.
@article{e288f63ca01a4c99af9dceef39d9cfa1,
title = "Design of DRAM-NAND flash hybrid main memory and Q-learning-based prefetching method",
abstract = "Owing to the increased need for machine learning and artificial intelligence in current cloud computing systems, the amount of data that needs to be processed has exponentially increased. Thus, it is important to optimize memory and storage systems to reduce the energy consumption and execution time of applications. This paper proposes a new Q-learning-based prefetching algorithm for DRAM–NAND flash hybrid main memory architecture. To minimize the computational overheads of learning-based schemes, we have designed two learning policies, namely aggressive learning and lazy learning. The proposed system reduces the energy consumption by about 80{\%} of the memory and storage for Redis, OpenStack Swift which is a cloud computing open source framework and Apache Storm workloads. Further, the overall execution time of workloads in cloud computing applications is reduced by almost half. Using a path generator with a Q-learning-based prefetching algorithm, we realize an increased hit rate of about 21{\%} compared to that with a no-prefetching system, compared to non-prefetching system.",
author = "Yoon, {Su Kyung} and Youn, {Young Sun} and Kim, {Jeong Geun} and Kim, {Shin Dug}",
year = "2018",
month = "10",
day = "1",
doi = "10.1007/s11227-018-2421-7",
language = "English",
volume = "74",
pages = "5293--5313",
journal = "Journal of Supercomputing",
issn = "0920-8542",
publisher = "Springer Netherlands",
number = "10",

}

Design of DRAM-NAND flash hybrid main memory and Q-learning-based prefetching method. / Yoon, Su Kyung; Youn, Young Sun; Kim, Jeong Geun; Kim, Shin Dug.

In: Journal of Supercomputing, Vol. 74, No. 10, 01.10.2018, p. 5293-5313.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Design of DRAM-NAND flash hybrid main memory and Q-learning-based prefetching method

AU - Yoon, Su Kyung

AU - Youn, Young Sun

AU - Kim, Jeong Geun

AU - Kim, Shin Dug

PY - 2018/10/1

Y1 - 2018/10/1

N2 - Owing to the increased need for machine learning and artificial intelligence in current cloud computing systems, the amount of data that needs to be processed has exponentially increased. Thus, it is important to optimize memory and storage systems to reduce the energy consumption and execution time of applications. This paper proposes a new Q-learning-based prefetching algorithm for DRAM–NAND flash hybrid main memory architecture. To minimize the computational overheads of learning-based schemes, we have designed two learning policies, namely aggressive learning and lazy learning. The proposed system reduces the energy consumption by about 80% of the memory and storage for Redis, OpenStack Swift which is a cloud computing open source framework and Apache Storm workloads. Further, the overall execution time of workloads in cloud computing applications is reduced by almost half. Using a path generator with a Q-learning-based prefetching algorithm, we realize an increased hit rate of about 21% compared to that with a no-prefetching system, compared to non-prefetching system.

AB - Owing to the increased need for machine learning and artificial intelligence in current cloud computing systems, the amount of data that needs to be processed has exponentially increased. Thus, it is important to optimize memory and storage systems to reduce the energy consumption and execution time of applications. This paper proposes a new Q-learning-based prefetching algorithm for DRAM–NAND flash hybrid main memory architecture. To minimize the computational overheads of learning-based schemes, we have designed two learning policies, namely aggressive learning and lazy learning. The proposed system reduces the energy consumption by about 80% of the memory and storage for Redis, OpenStack Swift which is a cloud computing open source framework and Apache Storm workloads. Further, the overall execution time of workloads in cloud computing applications is reduced by almost half. Using a path generator with a Q-learning-based prefetching algorithm, we realize an increased hit rate of about 21% compared to that with a no-prefetching system, compared to non-prefetching system.

UR - http://www.scopus.com/inward/record.url?scp=85047376536&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047376536&partnerID=8YFLogxK

U2 - 10.1007/s11227-018-2421-7

DO - 10.1007/s11227-018-2421-7

M3 - Article

AN - SCOPUS:85047376536

VL - 74

SP - 5293

EP - 5313

JO - Journal of Supercomputing

JF - Journal of Supercomputing

SN - 0920-8542

IS - 10

ER -