A batch scheduler with high level components

Capit, Nicolas; Costa, Georges Da; Georgiou, Yiannis; Huard, Guillaume; Martin, Cyrille; Mounié, Grégory; Neyron, Pierre; Richard, Olivier

doi:10.1109/ccgrid.2005.1558641

Cited by 149 publications

(130 citation statements)

References 5 publications

Supporting

Mentioning

129

Contrasting

Unclassified

Order By: Relevance

“…In [13] the viability of this approach was shown. Those images can be either built or downloaded from existing testbed infrastructures (e.g Grid'5000, FutureGrid) or sites as TURNKEY 4 or Cloud market 5 oriented to Amazon EC2 images. Those images are independent from the ones provided by the platform and experimenters have access to more operating system flavors.…”

Section: Methods For Setting Up the Environment Of Experimentationmentioning

confidence: 99%

“…In order to show that our approach is very portable between versions of Linux distributions. We carried out successfully construction and reconstruction of different appliances as shown in Table 1 that consist in different flavors of GNU/Linux (Debian, Ubuntu) and middleware: OAR [4] a very lightweight batch scheduler, Hadoop 11 and TAU 12 . It was possible to reproduce old environments of test back to 2009.…”

Section: Persistent Cache Mechanismmentioning

confidence: 99%

See 1 more Smart Citation

Reproducible Software Appliances for Experimentation

Ruiz

Richard

Emeras

2014

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering

View full text Add to dashboard Cite

Abstract. Experiment reproducibility is a milestone of the scientific method. Reproducibility of experiments in computer science would bring several advantages such as code re-usability and technology transfer. The reproducibility problem in computer science has been solved partially, addressing particular class of applications or single machine setups. In this paper we present our approach oriented to setup complex environments for experimentation, environments that require a lot of configuration and the installation of several software packages. The main objective of our approach is to enable the exact and independent reconstruction of a given software environment and the reuse of code. We present a simple and small software appliance generator that helps an experimenter to construct a specific software stack that can be deployed on different available testbeds.

show abstract

Section: Methods For Setting Up the Environment Of Experimentationmentioning

confidence: 99%

Section: Persistent Cache Mechanismmentioning

confidence: 99%

Reproducible Software Appliances for Experimentation

Ruiz

Richard

Emeras

2014

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering

View full text Add to dashboard Cite

show abstract

“…Petrucci [43] proposed a unified framework for large grid computing systems using different techniques such as policies for power off idle nodes, thereby predicting the future load on servers to save power. Capit [44] highlighted different aspects of developing a green IT implementation framework to build green data centers, but this proposal lacks different techniques to focus on the emission of greenhouse gases, proper recycling policy, and service-level agreements between service providers and end users. Based on these findings, this paper proposes a novel, measurement-based method to characterize and improve the power efficiency of a data center.…”

Section: Related Workmentioning

confidence: 99%

Power Usage Effectiveness Metrics to Measure Efficiency and Performance of Data Centers

Uddin

Alsaqour

Shah

et al. 2014

Appl. Math. Inf. Sci.

View full text Add to dashboard Cite

Abstract:Intensifying computation demand from enterprises has driven the growth of large, multifaceted data centers to manage current Internet, financial, commercial, and business applications. A data center comprises thousands of servers and other equipment that require substantial amounts of power to operate. This condition results in numerous challenges for the data center industry, such as massive energy consumption, underutilization of installed equipment, emission of greenhouse gases, and effect on global warming. This paper highlights the significance of identifying metrics to determine the performance and efficiency of a data center, which can help such a facility achieve operational cost savings through proper implementation of performance-measuring metrics. This paper discusses the implementation of Power Usage Effectiveness metrics in a tier-level data center in Pakistan. The results show that the overall performance value of the facility is 3.3, which indicates poor and inefficient operations.

show abstract

“…An improved support for moldability can be found in OAR [9]. The user gives a list of host-counts and walltimes, then the RMS chooses the configuration which minimizes the job's completion time.…”

Section: B Rms Support For Moldabilitymentioning

confidence: 99%

“…Thus, CooRM is compared to an RMS that gives equivalent schedules, so as to focus on the overhead that CooRM may bring. To this end, we have chosen OAR [9] both because it uses the CBF scheduling algorithm and due to its support for moldable jobs. In essence, each job is submitted with a list of host-count, wall-time pairs and OAR greedily chooses the configuration that minimizes the job's completion-time.…”

Section: A Overviewmentioning

confidence: 99%