Fair-share scheduling

About: Fair-share scheduling is a research topic. Over the lifetime, 24724 publications have been published within this topic receiving 516648 citations.

Biogeography-based optimization of makespan and reliability in grid computing systems

Abstract: The aim of this paper is to propose a scheduling method to consider reliability along with makespan in grid computing systems. The reliability of task execution is considered in the proposed method to increase the chance of running large-scale and computationally intensive workflows successfully. To handle situations in which a resource failure in a possible scheduling solution occurs, the proposed method finds a collection of scheduling solutions instead of only one solution to run the workflow. It leads to have chance to run an alternative scheduling solution that is not using the failed resource. To find the most optimized scheduling solutions, we exploit the lately developed biogeography-based optimization method with evaluation strategy and combine it with the operations like neighborhood search and crossover. Finally, the proposed method is compared with two successive scheduling methods. The results obtained from simulations show that gained improvement is significant especially in large workflows with large number of tasks.

Optimized Scheduling and Resource Allocation Using Evolutionary Algorithms in Cloud Environment

Abstract: Cloud computing is a powerful computing technology, which render a flexible services at anywhere to the user. One of the major issue of cloud computing was scheduling. In this work, a bacterial foraging optimization algorithm with genetic algorithm (GABFO) was combined to find out trustworthy scheduling problems in cloud workflow. Generally job scheduling and resource allocation in cloud is a tedious optimization problem at the time of considering QoS requirements. Lot of existing works under scheduling only concentrates on cost optimization and deadline problems, and it ignores the importance of reliability, availability and robustness. The main subscription of my work is to state a new optimized approach to schedule the jobs efficiently and allocate the resources in a efficient manner by introducing GABFO algorithm. Experiments were done in PSO, Genetic, BFO and then Genetic and BFO was combined to generate a hybrid optimized result, and the work was compared with above mentioned algorithms. The algorithms were executed for 52 iterations and totally 10 runs are calculated. The size of the job as well as virtual machines was varied for each iteration to calculate performance variation. We considered the optimization parameter as time and cost, and throughput. The work is implemented in cloudsim to create a simulated cloud environment. Final result shows better performance and maximum resource utilization in GABFO when compared to PSO, GA, BFO.

Minimizing Arbitrary Earliness/Tardiness Penalties with Common Due Date in Single-Machine Scheduling Problem Using a Tabu-Geno-Simulated Annealing

Abstract: In this paper, the single machine scheduling problem for a common due date with arbitrary earliness/tardiness penalties is discussed. The objective is to determine the common due date and processing sequence of new jobs together with the re-sequencing of old jobs to minimize the sum of total earliness/tardiness (E/T), completion time, and due date (dd) related penalties. We propose a new approach Tabu-Geno-Simulated Annealing (TGSA) by hybridization of three well-known metaheuristics, which have proven to be effective for this non deterministic polynomial time (NP) hard scheduling problem, namely, genetic algorithms (GA), tabu search, and simulated annealing (SA). Computational results have shown the effectiveness of the proposed approach in comparison with the ad-hoc heuristics on various single-machine scheduling problems. The assessment of the proposed hybridization indicates the efficiency in both the result and the time versus the other methods.

Stabilizing execution time of user processes by bottom half scheduling in Linux

Abstract: The CPU time allocated to user processes is rendered inaccurate by an unexpectedly and frequently occurring interrupt and a bottom half that consumes most interrupt processing time. Additionally, when the time consumed in the kernel mode greatly fluctuates with interrupt processing, the scheduler cannot distribute CPU time to user processes normally. This problem can dramatically distort the stable execution time of user processes. In addition, such time-sensitive applications as multimedia players cannot provide consistent quality. To overcome this stolen-time problem, we propose a bottom half scheduling approach that dynamically restricts the maximum time consumed by bottom halves. In this paper, we implement our proposed scheme in Linux 2.4. In addition, we show that the fluctuation of CPU time allocated to user processes by stolen-time can be shrunk with our proposed scheme by means of experiments using a multimedia application.