网格市场中服务质量驱动下的任务调度算法 |
| |
作者姓名: | 宋浒 杨寿保 刘晓茜 郭良敏 |
| |
作者单位: | 1. 中国科学技术大学计算机科学与技术学院, 合肥 230027;
2. 安徽师范大学计算机科学与技术系, 芜湖 241000 |
| |
基金项目: | 国家863项目(2006AA01A110)和国家自然科学基金(60273041)资助 |
| |
摘 要: | 针对资源提供方不能完成用户所有任务的情况,提出一种服务质量驱动下的任务调度算法,即预算和截止时间限制下的最大任务完成数调度算法(DBCN).这种批调度算法结合了Min-min算法吞吐量较高和线性规划全局优化的优点,不仅考虑了用户的所有任务,同时还考虑了优先级较高的任务.实验结果表明,该算法在任务完成总数方面比经典算法Min-min和DBCT分别提高了约10.6%和22.0%,在优先级高的任务完成总数方面也有大幅度提高,分别约为20%和40%.
|
关 键 词: | 网格市场 服务质量 预算和截止时间 调度算法 |
收稿时间: | 2010-03-05 |
修稿时间: | 2010-05-10 |
A quality-driven algorithm for task scheduling in grid market |
| |
Authors: | SONG Hu YANG Shou-Bao LIU Xiao-Qian GUO Liang-Min |
| |
Institution: | 1. School of Computer Science and Technology, University of Science and Technology of China, Hefei 230027, China;
2. Department of Computer Science and Technology, Anhui Normal University, Wuhu 241000, China |
| |
Abstract: | We propose a quality-driven algorithm for task scheduling in grid market, which is deadline- and budget-constrained and maximizes number of completed tasks (DBCN). This algorithm combines the high throughput advantage of Min-min algorithm and the global optimization advantage of linear programming. Meanwhile the algorithm considers not only all the tasks but also those prior ones. Compared with the Min-min and DBCT classical algorithms, DBCN completes about 10.6% and 22.0% more tasks and about 20% and 40% more prior tasks, respectively. |
| |
Keywords: | grid market QoS deadline and budget scheduling algorithm |
|
| 点击此处可从《》浏览原始摘要信息 |
| 点击此处可从《》下载免费的PDF全文 |
|