课题基金 / 基金详情

Kan: A Platform for Reliable Distributed Objects

Kan: A Platform for Reliable Distributed Objects
Kan:可靠的分布式对象平台
批准号:
9972571
负责人:
Ambuj Singh
金额:
$12.0万
依托单位国家:
美国
项目类别:
Standard Grant
财政年份:
1999
资助国家:
美国
项目状态:
已结题
起止时间:
1999-08-01 至 2001-12-31

项目摘要

项目成果

Ambuj Singh的其他基金

相似基金

相关文献

中文摘要
翻译
提案ccr -9972571题目:Kan:一个可靠的分布式对象平台spi: Ambuj K singh该项目的目标是设计和实现一个分布式面向对象系统,以支持大型和复杂的并行和分布式应用。该系统支持一个简单而强大的编程模型,具有以下语言特性:异步方法调用、并发线程、守卫、嵌套事务以及可以透明地迁移和复制的对象。该系统通过动态运行时优化(如自适应线程池、线程内联、对象访问模式监控、动态集群和粒度)实现了效率和可伸缩性。系统对节点故障具有弹性;这是通过低开销的日志和恢复算法实现的。研究允许处理器之间减少通信的较弱的对象一致性概念是该项目的另一个目标。共享白板、电子商务和并行科学项目是一些目标应用。该项目将有助于更好地理解可伸缩、容错的分布式系统。
英文摘要
Proposal CCR-9972571Title: Kan: A Platform for Reliable Distributed ObjectsPI: Ambuj K SinghThe goal of the project is to design and implementa distributed object-oriented system for supporting large and complex parallel and distributed applications. The system supports a simple, yet powerful, programming model with thefollowing language features: asynchronous method calls, concurrent threads, guards, nested transactions, and objects that can be migrated and replicated transparently. The system achieves efficiency and scalability through dynamic runtime optimizations such as an adaptive thread pool,thread inlining, monitoring of object access patterns,and dynamic clustering and granularity.The system is resilient to node failures; this is attainedthrough low-overhead algorithms for logging and recovery.Investigating weaker notions of object consistency that allow reduced communication between processorsis another goal of the project.Shared whiteboards, electroniccommerce, and parallel scientific programs are some of the targeted applications.The project will lead to a better understanding of scalable, fault-tolerant distributed systems.
期刊论文(0)
专著(0)
科研奖励(0)
会议论文
HDR DSC: Collaborative Research: Central Coast Data Science Partnership: Training a New Generation of Data Scientists
III: Small: Explaining heterogeneity within and across evolving networks
IGERT-CIF21: Interdisciplinary Graduate Education Research and Training in Network Science
III: Small: Modeling, Querying and Mining of Dynamic Graphs
国内基金
海外基金
Data-driven Recommendation System Construction of an Online Medical Platform Based on the Fusion of Information