Fairness and scheduling in single server queues

Wierman, Adam

doi:10.1016/j.sorms.2010.07.002

Cited by 52 publications

(27 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The second metric focuses on the notion of fairness. Even if there are different ways to measure fairness [33], a common approach is to consider the slowdown, i.e., the ratio between a job's response time and its size. In this work we focus on the per-job slowdown, to analyze whether jobs are treated unfairly, which translates into very high slowdown values.…”

Section: B Metricsmentioning

confidence: 99%

HFSP: Bringing Size-Based Scheduling To Hadoop

Pastorelli¹,

Carra

Dell’Amico³

et al. 2017

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

Abstract-Size-based scheduling with aging has been recognized as an effective approach to guarantee fairness and nearoptimal system response times. We present HFSP, a scheduler introducing this technique to a real, multi-server, complex and widely used system such as Hadoop.Size-based scheduling requires a priori job size information, which is not available in Hadoop: HFSP builds such knowledge by estimating it on-line during job execution.Our experiments, which are based on realistic workloads generated via a standard benchmarking suite, pinpoint at a significant decrease in system response times with respect to the widely used Hadoop Fair scheduler, without impacting the fairness of the scheduler, and show that HFSP is largely tolerant to job size estimation errors.

show abstract

Section: B Metricsmentioning

confidence: 99%

HFSP: Bringing Size-Based Scheduling To Hadoop

Pastorelli¹,

Carra

Dell’Amico³

et al. 2017

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

show abstract

“…A barrier toll is usually regarded as the bottleneck on highways, however we want to minimize the hindering effect to the traffic flow because of the barrier toll system [5][6][7]. So an ideal flashed into our mind that if all vehicles spend the same time through the barrier toll and have certain as well as similar speed back to the highway, then the barrier toll system can be ignored to every vehicle.…”

Section: A Avoid Congestionmentioning

confidence: 99%

Creative Design for Barrier Tolls

Liu¹,

Fei²,

Tao³

2017

Proceedings of the 2017 International Conference on Management, Education and Social Science (ICMESS 2017)

View full text Add to dashboard Cite

Abstract-This paper is aimed to design the toll plaza and its merging pattern to relieve the congestion as much as possible. Firstly, we calculate the necessary number of tollbooths for L traffic lanes to avoid jam in the fan-out area is 4L. Secondly, we want to decide the arranging shape of tollbooths. Compared traditional barrier toll with tollbooths, we are considering another shape to arrange tollbooths to avoid the shortcoming of traditional merging pattern that the overlapping area of neighboring vehicles' driving track often cause crash. Our model includes innovative and pioneering thinking and is especially suitable for autonomous vehicles, which means it will have extensive application in the future.

show abstract

“…A scheduling method can be selected based on the match between the requirements of the target environment and the goals of that method [3]. Scheduling has been mainly studied to minimize the turnaround time, waiting time, and response time, and to maximize CPU utilization, throughput [31], [34]. For instance, interactive systems are commonly used on personal computers and servers, besides optimizing response time is an important requirement and a primary goal in such environment [9], [19].…”

Section: Related Workmentioning

confidence: 99%

“…However, optimizing response time by minimizing measures other than variance in the response time may not necessarily indicate that all users get good service [13], [30], and cannot guarantee fairness or QoS among users [12], [31]. Some processes may get a priority at the expense of others [22] e.g.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On Minimum Variance CPU-Scheduling Algorithm for Interactive Systems using Goal Programming

Husain¹

2016

IJCA

View full text Add to dashboard Cite

Improving response time is considered a fundamental objective in interactive environments. CPU scheduling aimed mainly to optimize the response time by minimizing its average in order to attain faster responses to users' requests. However, for interactive systems, reasonable and predictable services are more preferred than faster responses but highly variable. Delivering service in a timely manner at less variable response time is an issue that has been addressed in this paper. A goal programming (GP) model is proposed to perform CPU scheduling at minimum variance and low response time. The GP method determines the optimal process in the ready queue that best minimizes the variance to be executed first. A simulation system that can generate varied scheduling situations was developed and several tests were conducted. The performance of the proposed GP scheduling method is measured and compared to the other related scheduling methods. The evaluation results show that the GP scheduling method can provide predictable and reasonable service and it performs scheduling at minimum variance and lower response time. The GP method outperforms the other related methods with varying degrees.

show abstract

Fairness and scheduling in single server queues

Cited by 52 publications

References 45 publications

HFSP: Bringing Size-Based Scheduling To Hadoop

HFSP: Bringing Size-Based Scheduling To Hadoop

Creative Design for Barrier Tolls

On Minimum Variance CPU-Scheduling Algorithm for Interactive Systems using Goal Programming

Contact Info

Product

Resources

About