Data cache management using frequency-based replacement

John T. Robinson, Murthy Devarakonda

Research output: Chapter in Book/Report/Conference proceedingConference contribution

167 Citations (Scopus)

Abstract

We propose a new frequency-based replacement algorithm for managing caches used for disk blocks by a file system, database management system, or disk control unit, which we refer to here as data caches. Previously, LRU replacement has usually been used for such caches. We describe a replacement algorithm based on the concept of maintaining reference counts in which locality has been 'factored out'. In this algorithm replacement choices are made using a combination of reference frequency and block age. Simulation results based on traces of file system and I/O activity from actual systems show that this algorithm can offer up to 34% performance improvement over LRU replacement, where the improvement is expressed as the fraction of the performance gain achieved between LRU replacement and the theoretically optimal policy in which the reference string must be known in advance. Furthermore, the implementation complexity and efficiency of this algorithm is comparable to one using LRU replacement.

Original languageEnglish (US)
Title of host publication1990 ACM SIGMETRICS Conf Meas Model Comput Syst
Editors Anon
PublisherPubl by ACM
Pages134-142
Number of pages9
ISBN (Print)0897913590
StatePublished - Dec 1 1990
Externally publishedYes
Event1990 ACM SIGMETRICS Conference on Measurement & Modeling of Computer Systems - Boulder, CO, USA
Duration: May 22 1990May 25 1990

Other

Other1990 ACM SIGMETRICS Conference on Measurement & Modeling of Computer Systems
CityBoulder, CO, USA
Period5/22/905/25/90

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Robinson, J. T., & Devarakonda, M. (1990). Data cache management using frequency-based replacement. In Anon (Ed.), 1990 ACM SIGMETRICS Conf Meas Model Comput Syst (pp. 134-142). Publ by ACM.

Data cache management using frequency-based replacement. / Robinson, John T.; Devarakonda, Murthy.

1990 ACM SIGMETRICS Conf Meas Model Comput Syst. ed. / Anon. Publ by ACM, 1990. p. 134-142.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Robinson, JT & Devarakonda, M 1990, Data cache management using frequency-based replacement. in Anon (ed.), 1990 ACM SIGMETRICS Conf Meas Model Comput Syst. Publ by ACM, pp. 134-142, 1990 ACM SIGMETRICS Conference on Measurement & Modeling of Computer Systems, Boulder, CO, USA, 5/22/90.
Robinson JT, Devarakonda M. Data cache management using frequency-based replacement. In Anon, editor, 1990 ACM SIGMETRICS Conf Meas Model Comput Syst. Publ by ACM. 1990. p. 134-142
Robinson, John T. ; Devarakonda, Murthy. / Data cache management using frequency-based replacement. 1990 ACM SIGMETRICS Conf Meas Model Comput Syst. editor / Anon. Publ by ACM, 1990. pp. 134-142
@inproceedings{d06b6629439c45de95fe4c47451bcee7,
title = "Data cache management using frequency-based replacement",
abstract = "We propose a new frequency-based replacement algorithm for managing caches used for disk blocks by a file system, database management system, or disk control unit, which we refer to here as data caches. Previously, LRU replacement has usually been used for such caches. We describe a replacement algorithm based on the concept of maintaining reference counts in which locality has been 'factored out'. In this algorithm replacement choices are made using a combination of reference frequency and block age. Simulation results based on traces of file system and I/O activity from actual systems show that this algorithm can offer up to 34{\%} performance improvement over LRU replacement, where the improvement is expressed as the fraction of the performance gain achieved between LRU replacement and the theoretically optimal policy in which the reference string must be known in advance. Furthermore, the implementation complexity and efficiency of this algorithm is comparable to one using LRU replacement.",
author = "Robinson, {John T.} and Murthy Devarakonda",
year = "1990",
month = "12",
day = "1",
language = "English (US)",
isbn = "0897913590",
pages = "134--142",
editor = "Anon",
booktitle = "1990 ACM SIGMETRICS Conf Meas Model Comput Syst",
publisher = "Publ by ACM",

}

TY - GEN

T1 - Data cache management using frequency-based replacement

AU - Robinson, John T.

AU - Devarakonda, Murthy

PY - 1990/12/1

Y1 - 1990/12/1

N2 - We propose a new frequency-based replacement algorithm for managing caches used for disk blocks by a file system, database management system, or disk control unit, which we refer to here as data caches. Previously, LRU replacement has usually been used for such caches. We describe a replacement algorithm based on the concept of maintaining reference counts in which locality has been 'factored out'. In this algorithm replacement choices are made using a combination of reference frequency and block age. Simulation results based on traces of file system and I/O activity from actual systems show that this algorithm can offer up to 34% performance improvement over LRU replacement, where the improvement is expressed as the fraction of the performance gain achieved between LRU replacement and the theoretically optimal policy in which the reference string must be known in advance. Furthermore, the implementation complexity and efficiency of this algorithm is comparable to one using LRU replacement.

AB - We propose a new frequency-based replacement algorithm for managing caches used for disk blocks by a file system, database management system, or disk control unit, which we refer to here as data caches. Previously, LRU replacement has usually been used for such caches. We describe a replacement algorithm based on the concept of maintaining reference counts in which locality has been 'factored out'. In this algorithm replacement choices are made using a combination of reference frequency and block age. Simulation results based on traces of file system and I/O activity from actual systems show that this algorithm can offer up to 34% performance improvement over LRU replacement, where the improvement is expressed as the fraction of the performance gain achieved between LRU replacement and the theoretically optimal policy in which the reference string must be known in advance. Furthermore, the implementation complexity and efficiency of this algorithm is comparable to one using LRU replacement.

UR - http://www.scopus.com/inward/record.url?scp=0025537206&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0025537206&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0025537206

SN - 0897913590

SP - 134

EP - 142

BT - 1990 ACM SIGMETRICS Conf Meas Model Comput Syst

A2 - Anon, null

PB - Publ by ACM

ER -