Improvements in microprocessor and networking performance have made cluster system a very attractive platform for high-end parallel and distributed computing 随着微处理器技术和高性能网络技术的飞速发展,集群计算逐渐成为一种有成本效益的并行分布式计算资源。
After discussing the strategies of data migration and scalability, we propose a dynamic data migrating algorithm comparing with disk cache in the clustered system 本文讨论了它的数据迁移策略和系统扩展性问题,并与集群式磁盘cache进行比较,提出了一种动态数据迁移算法。
Because lsf is a quite mature commercial end product software, therefore we have designed based on the lsf heterogeneity multi-clusters system . in the system has 本文设计了基于lsf成品商业软件的异构多机群系统,系统由lsf主机群、sge子机群、pbs子机群、lsf子机群共同构成。
The cluster system maybe consists of heterogeneous computer system, but the current parallel programming environment ca n't cross the different operate systems to execute parallel computing 机群系统具有异构性,现有的并行编程环境存在不能跨越异构操作系统的问题,不能跨平台进行并行计算。
This thesis constructed a runable virtual server cluster system, and implemented the high reliability virtual server load dispatching scheme on real-time embed operating system vxworks 本文构建了一个可运行的虚拟服务集群系统,在嵌入式实时操作系统vxworks上实现了高可靠性的虚拟服务负载分发方案。
The special replacemetlt algorithin of this information cache assures that the nodes in the cache are the k most overloaded nodes and the k most underloaded nodes in the cluster system 信息cache通过特殊的替换算法,保证两个cache中的信息分别为系统中负载最大的k个超载节点的信息和负载最小的k个欠载节点的信息。
Currently, web server cluster system has been adopted as a main solution by many web service providers, because it has much more technical advantages than single web server 相对于单台web服务器,web服务器集群系统在技术上具有明显的不可比拟的优势,因而已被众多web服务提供商采纳作为主要的web服务器解决方案。
4 . to deal with the runtime performance management task of component cluster system, we firstly discuss the optimizing technologies of single component application server, which includes memory optimization, connection mana 在此基础上,本文实现了负载均衡调度管理设施ha一imr及支持企业高可用计算需求的集群系统结点管理设施ha一oa_hb_lmd。
In this dissertation, research has been made on the topics of clock of cluster system, time-cost of debugging, intrusion of interactive debugging operation, dynamic conditional breakpoint, and performance debugging issues 本文对并行调试中的时钟模型、时间开销、交互式调试的干扰、动态条件断点、性能调试等问题进行了研究,取得了有价值的成果。
In this paper, we introduced briefly the technology of cluster system, the proposal and development of the tools for performance analysis and monitor, the classification and current developing situation of the common performance analysis tools 本论文简要介绍了机群系统技术、性能分析与监测工具的提出和发展、常见性能分析工具分类及其当前研究现状。