Mining for Misconfigured Machines in Grid Systems

Noam Palatin; Arie Leizarowitz; Assaf Schuster; Ran Wolff

doi:10.1002/9780470699904.ch5

Mining for Misconfigured Machines in Grid Systems

Noam Palatin, Arie Leizarowitz, Assaf Schuster, Ran Wolff

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

Abstract

This chapter describes the Grid Monitoring System (GMS) - a system which adopts a distributed data mining approach to detection of misconfigured grid machines. The GMS non-intrusively collects data from sources available throughout the grid system. It converts raw data to semantically meaningful data and stores these data on the machine from which, it was obtained limiting incurred overhead and allowing scalability. When analysis is requested, a distributed outliers detection algorithm is employed to identify misconfigured machines. The algorithm itself is implemented as a recursive workflow of grid jobs and is especially suited to grid systems in which the machines might be unavailable most of the time or often fail altogether.

Original language	English
Title of host publication	Data Mining Techniques in Grid Computing Environments
Publisher	John Wiley & Sons, Ltd
Pages	71-89
Number of pages	19
ISBN (Print)	9780470512586
DOIs	https://doi.org/10.1002/9780470699904.ch5
State	Published - 22 Jun 2009
Externally published	Yes

Keywords

Data analysis
Distributed HilOut algorithm
GMS ontology - a job
GMS using Condor DAGman
Grid Monitoring System (GMS)
Grid system misconfigured machine mining
Knowledge discovery process - acquiring
Large grid system misconfigured machine detection
Pre-processing and storing data
System data acquisition - intrusive and non-intrusive
a pool and matchmaking

ASJC Scopus subject areas

General Computer Science

Access to Document

10.1002/9780470699904.ch5

Cite this

@inbook{fa69881dcc4d451ba526eec9e3e20f32,

title = "Mining for Misconfigured Machines in Grid Systems",

abstract = "This chapter describes the Grid Monitoring System (GMS) - a system which adopts a distributed data mining approach to detection of misconfigured grid machines. The GMS non-intrusively collects data from sources available throughout the grid system. It converts raw data to semantically meaningful data and stores these data on the machine from which, it was obtained limiting incurred overhead and allowing scalability. When analysis is requested, a distributed outliers detection algorithm is employed to identify misconfigured machines. The algorithm itself is implemented as a recursive workflow of grid jobs and is especially suited to grid systems in which the machines might be unavailable most of the time or often fail altogether.",

keywords = "Data analysis, Distributed HilOut algorithm, GMS ontology - a job, GMS using Condor DAGman, Grid Monitoring System (GMS), Grid system misconfigured machine mining, Knowledge discovery process - acquiring, Large grid system misconfigured machine detection, Pre-processing and storing data, System data acquisition - intrusive and non-intrusive, a pool and matchmaking",

author = "Noam Palatin and Arie Leizarowitz and Assaf Schuster and Ran Wolff",

year = "2009",

month = jun,

day = "22",

doi = "10.1002/9780470699904.ch5",

language = "English",

isbn = "9780470512586",

pages = "71--89",

booktitle = "Data Mining Techniques in Grid Computing Environments",

publisher = "John Wiley & Sons, Ltd",

}

TY - CHAP

T1 - Mining for Misconfigured Machines in Grid Systems

AU - Palatin, Noam

AU - Leizarowitz, Arie

AU - Schuster, Assaf

AU - Wolff, Ran

PY - 2009/6/22

Y1 - 2009/6/22

N2 - This chapter describes the Grid Monitoring System (GMS) - a system which adopts a distributed data mining approach to detection of misconfigured grid machines. The GMS non-intrusively collects data from sources available throughout the grid system. It converts raw data to semantically meaningful data and stores these data on the machine from which, it was obtained limiting incurred overhead and allowing scalability. When analysis is requested, a distributed outliers detection algorithm is employed to identify misconfigured machines. The algorithm itself is implemented as a recursive workflow of grid jobs and is especially suited to grid systems in which the machines might be unavailable most of the time or often fail altogether.

AB - This chapter describes the Grid Monitoring System (GMS) - a system which adopts a distributed data mining approach to detection of misconfigured grid machines. The GMS non-intrusively collects data from sources available throughout the grid system. It converts raw data to semantically meaningful data and stores these data on the machine from which, it was obtained limiting incurred overhead and allowing scalability. When analysis is requested, a distributed outliers detection algorithm is employed to identify misconfigured machines. The algorithm itself is implemented as a recursive workflow of grid jobs and is especially suited to grid systems in which the machines might be unavailable most of the time or often fail altogether.

KW - Data analysis

KW - Distributed HilOut algorithm

KW - GMS ontology - a job

KW - GMS using Condor DAGman

KW - Grid Monitoring System (GMS)

KW - Grid system misconfigured machine mining

KW - Knowledge discovery process - acquiring

KW - Large grid system misconfigured machine detection

KW - Pre-processing and storing data

KW - System data acquisition - intrusive and non-intrusive

KW - a pool and matchmaking

UR - http://www.scopus.com/inward/record.url?scp=84889970718&partnerID=8YFLogxK

U2 - 10.1002/9780470699904.ch5

DO - 10.1002/9780470699904.ch5

M3 - Chapter

AN - SCOPUS:84889970718

SN - 9780470512586

SP - 71

EP - 89

BT - Data Mining Techniques in Grid Computing Environments

PB - John Wiley & Sons, Ltd

ER -

Mining for Misconfigured Machines in Grid Systems

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this