| 引用本文: | 李鑫,杨学军,徐新海.分布式流体系结构及其编程模型与资源管理.[J].国防科技大学学报,2015,37(6):110-115.[点击复制] |
| LI Xin,YANG Xuejun,XU Xinhai.Programming model and resource management of distributed stream architecture[J].Journal of National University of Defense Technology,2015,37(6):110-115[点击复制] |
|
| |
|
|
| 本文已被：浏览 8708次 下载 6780次 |
| 分布式流体系结构及其编程模型与资源管理 |
| 李鑫1,2,3, 杨学军1,2, 徐新海1,2 |
|
(1.国防科技大学 计算机学院， 湖南 长沙 410073;2.
2.国防科技大学 高性能计算国家重点实验室， 湖南 长沙 410073;3.
3.中国人民解放军总参谋部第六十三研究所， 江苏 南京 210007)
|
| 摘要: |
| 利用互联网资源提供大数据计算服务面临着资源异构性、动态性与通信长延迟等方面的挑战，现有分布式计算模型仍存在一些不足。运用流计算模型提出分布式流体系结构，包括分布式流编程模型与资源管理等，能够高效支持多种并行执行模式。在10个CPU-GPU异构结点上实现了原型系统，仿真实验验证了7个不同的测试用例。实验结果表明，与本地串行计算相比，分布式流体系结构可以平均提高39倍计算性能，具有较大的应用潜力。 |
| 关键词: 流体系结构 大数据 编程模型 分布式计算 |
| DOI：10.11887/j.cn.201506021 |
| 投稿日期：2015-09-06 |
| 基金项目:国家自然科学基金资助项目（61221491，61303071） |
|
| Programming model and resource management of distributed stream architecture |
| LI Xin1,2,3, YANG Xuejun1,2, XU Xinhai1,2 |
|
(1. College of Computer, National University of Defense Technology, Changsha 410073, China;2.
2. State Key Laboratory of High Performance Computing, National University of Defense Technology, Changsha 410073, China;3.
3. The 63rd Research Institute of PLA General Staff Headquarters, Nanjing 210007, China)
|
| Abstract: |
| While providing big data computing services using Internet resources, there remains a big challenge to researchers, including heterogeneity of Internet resources, dynamics of Internet resources and long latency of Internet communication. Current influent distributed computing models still have some shortage. A novel distributed stream computing model was proposed based on the traditional stream computing model, including the distributed stream programming model and resource management can efficiently support multiple parallel execution modes. The prototype system implemented on the 10 CPU-GPU heterogeneous nodes. Seven different benchmarks used in the simulation experiment. The experimental result shows that the distributed stream architecture can achieve the speedup of at least on average over the local serial computing, with significant potential for applications. |
| Keywords: stream architecture big data programming model distributed computing |
|
|
|
|
|
