生物计算网格中的在线调度技术研究  

BioLab : a bioinformatics grid computing system based on online scheduling

作　　者：刘文懋[1] 张伟哲[1] 张宏莉[1] 方滨兴[1] 

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,哈尔滨150001

出　　处：《高技术通讯》2009年第10期1014-1019,共6页Chinese High Technology Letters

基　　金：863计划(2006AA02Z334;2009AA01Z437);国家自然科学基金(60703014);973计划(G2005CB321806);高等学校博士学科点专项科研基金(20070213044);中国博士后科学基金(20070410263)资助项目

摘　　要：研究了在异构网格环境下的生物应用集成,定义了服务的提供者、部署者和使用者三种用户角色,设计了网格环境下的服务和资源整合机制,重点实现了用户管理以及作业调度控制等功能。根据计算资源的异构特点,设计了多种启发式调度算法。考虑到生物计算应用的不同类型,提出了自适应调度算法,该算法根据应用的特点动态选择启发式调度算法。实验表明,非阻塞调度优于阻塞调度方式;自适应调度算法比静态的在线调度算法有更好的性能,而在异构的网络中,带宽优先调度算法的性能比其他静态调度算法性能更好。In this paper bioinformatics application integration in the heterogeneous grid environment is studied, and the BioLab, a bioinformatics grid computing system based on online scheduling, is designed. Three roles are defined in the system： service providers, service distributers and service users. The integrated mechanism for services and resources is designed, and the modules such as the user management, the job scheduling and the job management are implemented. Several heuristic scheduling algorithms are designed according to the heterogeneous features of computing resources. The adaptive scheduling algorithm, which chooses heuristic scheduling algorithms dynamically based on alterative application characteristics, is proposed after considering different types of biological applications. The experiments show that in the large average job arrival time, unblocked scheduling is better than blocked scheduling, and the adaptive scheduling algorithm outperforms the static scheduling algorithms, while in the heterogeneous grid the bandwidth-first scheduling algorithm outperforms the other static scheduling algorithms.

关 键 词：生物信息学 计算网格 在线调度 BioLab 

分 类 号：TP399-C8[自动化与计算机技术—计算机应用技术]