Accelerating Linux/Windows File Systems by Predicting Access Frequency

Frank Wang; Chenhan Liao; Na Helian; chris Thompson; Sining Wu; Yuhui Deng; Vineet Khare; Michael Andrew Parker

Accelerating Linux/Windows File Systems by Predicting Access Frequency

Frank Wang, Chenhan Liao, Na Helian, chris Thompson, Sining Wu, Yuhui Deng, Vineet Khare, Michael Andrew Parker

Research output: Contribution to conference › Paper › peer-review

Abstract

Network File System (NFS, de facto in Linux) or Common Internet File System (CIFS, de facto in Windows) is used for applications to move files across networks/grids. We prove that once a file is created with a set of attributes, such as name, type, permission mode, owner and owner group, its future access frequency is predictable. A decision-tree-based predictive model is established with an integrated fuzzy logic facility to further calculate the degree to which a file may be frequently accessed. By consulting with the rules generated from the predictive model over diverse real-system NFS traces, it can predict a newly created file’s future access frequency with a sufficient accuracy (greater than 90%).

Original language	English
Pages	1-8
Number of pages	8
Publication status	Published - 13 Sept 2007
Event	the UK e-Science 2007 All Hands Meeting - Nottingham, United Kingdom Duration: 10 Sept 2007 → 13 Sept 2007 http://web.archive.org/web/20080227100209/http://www.allhands.org.uk/2007/

Conference

Conference	the UK e-Science 2007 All Hands Meeting
Country/Territory	United Kingdom
City	Nottingham
Period	10/09/07 → 13/09/07
Internet address	http://web.archive.org/web/20080227100209/http://www.allhands.org.uk/2007/

Cite this

@conference{85b59ab1dafa48d2992933c2c8997f06,

title = "Accelerating Linux/Windows File Systems by Predicting Access Frequency",

abstract = "Network File System (NFS, de facto in Linux) or Common Internet File System (CIFS, de facto in Windows) is used for applications to move files across networks/grids. We prove that once a file is created with a set of attributes, such as name, type, permission mode, owner and owner group, its future access frequency is predictable. A decision-tree-based predictive model is established with an integrated fuzzy logic facility to further calculate the degree to which a file may be frequently accessed. By consulting with the rules generated from the predictive model over diverse real-system NFS traces, it can predict a newly created file{\textquoteright}s future access frequency with a sufficient accuracy (greater than 90%).",

author = "Frank Wang and Chenhan Liao and Na Helian and chris Thompson and Sining Wu and Yuhui Deng and Vineet Khare and Parker, {Michael Andrew}",

year = "2007",

month = sep,

day = "13",

language = "English",

pages = "1--8",

note = "the UK e-Science 2007 All Hands Meeting ; Conference date: 10-09-2007 Through 13-09-2007",

url = "http://web.archive.org/web/20080227100209/http://www.allhands.org.uk/2007/",

}

TY - CONF

T1 - Accelerating Linux/Windows File Systems by Predicting Access Frequency

AU - Wang, Frank

AU - Liao, Chenhan

AU - Helian, Na

AU - Thompson, chris

AU - Wu, Sining

AU - Deng, Yuhui

AU - Khare, Vineet

AU - Parker, Michael Andrew

PY - 2007/9/13

Y1 - 2007/9/13

N2 - Network File System (NFS, de facto in Linux) or Common Internet File System (CIFS, de facto in Windows) is used for applications to move files across networks/grids. We prove that once a file is created with a set of attributes, such as name, type, permission mode, owner and owner group, its future access frequency is predictable. A decision-tree-based predictive model is established with an integrated fuzzy logic facility to further calculate the degree to which a file may be frequently accessed. By consulting with the rules generated from the predictive model over diverse real-system NFS traces, it can predict a newly created file’s future access frequency with a sufficient accuracy (greater than 90%).

AB - Network File System (NFS, de facto in Linux) or Common Internet File System (CIFS, de facto in Windows) is used for applications to move files across networks/grids. We prove that once a file is created with a set of attributes, such as name, type, permission mode, owner and owner group, its future access frequency is predictable. A decision-tree-based predictive model is established with an integrated fuzzy logic facility to further calculate the degree to which a file may be frequently accessed. By consulting with the rules generated from the predictive model over diverse real-system NFS traces, it can predict a newly created file’s future access frequency with a sufficient accuracy (greater than 90%).

UR - https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=6ceadbae1c2ab68a1610d5c9723a02ed533b5f01

UR - https://citeseerx.ist.psu.edu/doc_view/pid/6ceadbae1c2ab68a1610d5c9723a02ed533b5f01

M3 - Paper

SP - 1

EP - 8

T2 - the UK e-Science 2007 All Hands Meeting

Y2 - 10 September 2007 through 13 September 2007

ER -

Accelerating Linux/Windows File Systems by Predicting Access Frequency

Abstract

Conference

Other files and links

Fingerprint

Cite this