Toward balanced and sustainable job scheduling for production supercomputers

Tang, Wei; Ren, Dongxu; Lan, Zhiling; Desai, Narayan

doi:10.1016/j.parco.2013.08.007

Cited by 10 publications

(8 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Jobs running on HESs can be calculated by determining the number jobs at time T, on service queue of , which is divided by the service rate of at time T. Job of all HESs in a data center is calculated as follows: (8) Execution time of HES: (9) Execution time of all HESs: (10) Standard deviation of job: (11) JOB SCHEDULING DECISION After applying the jobs scheduling using the proposed technique, load balancing at run time will come in action. Load balancing will migrate some jobs of heavily loaded HESs to underloaded HESs.…”

Section: Job On Hessmentioning

confidence: 99%

See 1 more Smart Citation

Multiobjective Artificial Bee Colony based Job Scheduling for Cloud Computing Environment

Sethi¹,

Singh²,

Singh³

2018

IJMSC

View full text Add to dashboard Cite

Cloud computing has become the hottest issue due to its wide range of services. Due to a large number of users, it becomes more significant to provide high availability of services to cloud users. The majority of existing scheduling techniques in the cloud environment is NP-Complete in nature. Many researchers have utilized meta-heuristic techniques to schedule the jobs in cloud data centers. The majority of existing techniques such as Genetic Algorithm, Ant colony optimization, Non-dominated Sorting Genetic Algorithm (NSGA-III), etc. suffer from poor convergence speed. Also, most of these techniques are either based upon scheduling or load balancing. Therefore, to overcome these issues, a new Variance Honey Bee Behavior with multi-objective optimization method (VHBBMO) is proposed in this paper. Extensive experiments have been conducted by considering the various set of jobs. The experimental results have shown that the proposed method provides more significant results than available methods.

show abstract

Section: Job On Hessmentioning

confidence: 99%

“…A parallel job scheduling method is designed to schedule the parallel workload [9]. The list scheduling based server assignment technique is developed which always evaluate an optimal server assignment that minimizes the makespan [10].…”

Section: Introductionmentioning

confidence: 99%

Multiobjective Artificial Bee Colony based Job Scheduling for Cloud Computing Environment

Sethi¹,

Singh²,

Singh³

2018

IJMSC

View full text Add to dashboard Cite

show abstract

“…As mentioned in Schroeder & Harchol-Balter [1], finding a good rule for assigning jobs to host machines remains an open question at many supercomputing sites. Over the last decades, many job assignment rules across a variety of distributed computational resources have been studied, e.g., Schaar & Efe [2], Chlamtac et al [3], Feitelson et al [4], von Laszewski [5], Piro et al [6], Tang et al [7], etc.…”

Section: Introductionmentioning

confidence: 99%

“…Tang et al [7] studied scheduling policies for balancing workload. Downey [8] investigated the decision making on job assignment to a space-sharing parallel computer based on observed workloads.…”

Section: Introductionmentioning

confidence: 99%

A Queueing Analysis for Job Assignment on Two-Type Heterogeneous Supercomputer System

Wang¹,

Chen²,

Hwang³

2015

Proceedings of the International Conference on Computer Information Systems and Industrial Applications

View full text Add to dashboard Cite

Abstract-This paper introduces a queueing analysis of a two-tier service supercomputer system, where one type of processors offers service with a finite waiting buffer and the other offers unlimited waiting buffer. From the managerial viewpoint of system, a two-tier queueing model is developed to investigate the dynamic performance of the supercomputer system under finite buffer control. The queueing system is formulated as a state dependent quasi-birth-and-death (QBD) process with twodimensional state space. System performance measures can be obtained through the matrix geometric solution for such a QBD process.

show abstract

“…The simplest scheduling policy of rigid jobs is based on first come first served (FCFS), whose principle is to queue jobs in order of submission, and the queue head job will be served when spare processors can satisfy its requirement. However, FCFS has proved its poor performance due to generating resource fragmentation [2,3]. To overcome this problem, several backfilling policies [4,5,6] have been proposed and implemented in production schedulers.…”

mentioning

confidence: 99%

HRF

Tuo

Jin

et al. 2015

Proceedings of the 12th ACM International Conference on Computing Frontiers

View full text Add to dashboard Cite

Moldable jobs, which allow the number of allocated processors to be adjusted before running in clusters, have attracted increasing concern in parallel job scheduling research. Compared with traditional rigid jobs where the number of allocated processors is fixed, moldable jobs are more flexible and therefore have more potential for improving their average turnaround time (a crucial metric to describe performance of jobs in a cluster). Average turnaround time of moldable jobs depends greatly on resource allocation schemes. Unfortunately, existing schemes do not perform well in reducing average turnaround time, either because they only consider a single job's turnaround time instead of the average turnaround time of all jobs, or because they just aim at fairness between short and long jobs instead of their average turnaround time. In this paper, we investigate how resource allocation affects the average turnaround time of moldable jobs in clusters, and propose a scheme named HRF (highest revenue first), which allocates processors according to the highest revenue of shortening runtime. In our simulations, experimental results show that HRF can reduce average turnaround time up to 71% when compared with state-of-the-art schemes.

show abstract

Toward balanced and sustainable job scheduling for production supercomputers

Cited by 10 publications

References 17 publications

Multiobjective Artificial Bee Colony based Job Scheduling for Cloud Computing Environment

Multiobjective Artificial Bee Colony based Job Scheduling for Cloud Computing Environment

A Queueing Analysis for Job Assignment on Two-Type Heterogeneous Supercomputer System

HRF

Contact Info

Product

Resources

About