EAGER: Network-Constrained Distributed Primitives

EAGER：网络约束的分布式原语

• 批准号: 1059540
• 负责人: Nitin Vaidya
• 金额: $ 18.18万
• 依托单位: University of Illinois at Urbana-Champaign
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-09-01 至 2013-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1059540&HistoricalAwards=false
• 关键词: EAGER; Network; Constrained; Distributed; Primitives

Distributed applications require coordination between multiple nodes in a network, and they need to continue to operate correctly despite node failures. Implementation of such applications benefits from the availability of distributed primitives for commonly used distributed operations, such as clock synchronization or consensus. While most of the past distributed applications have traditionally used wired networks, there is an emerging trend towards the use of wireless networks for distributed applications.The focus of this interdisciplinary project is on design, analysis and implementation of distributed primitives for wireless networks. In wireless networks, the network capacity is often quite limited. This project explores the design of efficient algorithms for distributed primitives under network capacity constraints, while exploiting wireless network capabilities such as local broadcast and physical layer adaptation. The scope of the project includes network-constrained distributed primitives for synchronization, consensus and function computation. Expected results from the project include theoretical bounds on performance of distributed primitives under network capacity constraints, design of efficient algorithms for the primitives, and practical implementations based on insights gained from the theory. Outcomes from the project are expected to allow efficient implementation of future distributed applications in resource-constrained wireless networks. The project brings together networking and distributed systems viewpoints, and is expected to yield fundamental new insights of interest to both areas. The research outcomes from this project will be disseminated via publications in technical conferences and journals, and incorporated into distributed computing and wireless networking courses.

分布式应用程序需要在网络中的多个节点之间进行协调，并且需要在节点发生故障时继续正常运行。这类应用程序的实现受益于常用分布式操作（如时钟同步或一致性）的分布式原语的可用性。虽然过去的大多数分布式应用程序传统上使用有线网络，但在分布式应用程序中使用无线网络的趋势正在出现。这个跨学科项目的重点是无线网络分布式原语的设计、分析和实现。在无线网络中，网络容量通常是非常有限的。本项目探索在网络容量限制下分布式原语的高效算法设计，同时利用无线网络功能，如本地广播和物理层适应。该项目的范围包括用于同步、共识和函数计算的网络约束分布式原语。该项目的预期结果包括网络容量约束下分布式原语性能的理论界限、原语高效算法的设计以及基于从理论中获得的见解的实际实现。该项目的成果有望在资源受限的无线网络中有效实现未来的分布式应用。该项目将网络和分布式系统的观点结合在一起，并期望产生对这两个领域都感兴趣的基本的新见解。该项目的研究成果将通过技术会议和期刊的出版物传播，并纳入分布式计算和无线网络课程。