CAREER: Ensuring Sustainable Scalability for Globally-Distributed Systems

职业：确保全球分布式系统的可持续可扩展性

• 批准号: 0644062
• 负责人: Fabian Bustamante
• 金额: $ 40万
• 依托单位: Northwestern University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2007
• 资助国家: 美国
• 起止时间: 2007-01-15 至 2012-03-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0644062&HistoricalAwards=false
• 关键词: CAREER; Ensuring; Sustainable; Scalability; Globally

The past few years have witnessed a growing number of large-scalenetworked systems. Most of these systems are built following anoverlay approach, with each of them regularly and independentlyprobing its environment to guide path selection algorithms, routearound faulty links and replicate content for availability. As thesesystems grow in popularity, such an approach will result in anunsustainable degree of monitoring and restrict the variety, numberand span of distributed services.The thesis of this project is that a large fraction ofglobally-distributed systems can be built to ensure sustainablescalability by strategically reusing the view of the network gatheredby long-running, ubiquitous services such as CDNs and P2Psystems. This work defines and explores "3R" - a new approach to thedesign and implementation of distributed systems that focuses onminimizing aggregated control and administrative overhead bystrategically reusing environment's views and recycling previouslygathered measurements. In particular, we are designing efficienttechniques for maintaining, accessing and reusing this information forbuilding next-generation streaming multicast, content distribution anddata sharing applications.The work explores the tradeoffs in the recycling/reuse ofenvironmental measurements, and their implication indistributed-system design. Ultimately, it will facilitate deploymentsof wide-area applications, benefiting society at large. The researchagenda is complemented by a thorough education and outreach plan forstrengthening experimental systems education and contributing tominority recruitment and retention.

在过去的几年里，大规模网络系统的数量不断增加。这些系统中的大多数都是按照覆盖方法构建的，每个系统都定期独立地探测其环境，以指导路径选择算法，绕过故障链路并复制内容以确保可用性。随着这些系统的普及，这种方法将导致一个不可持续的程度的监控和限制的品种，数量和跨度的分布式services.The论文的这个项目是，一个大部分的全球分布式系统可以建立，以确保可持续的可扩展性，通过战略性地重用的网络视图收集的长期运行，无处不在的服务，如CDN和P2P系统。这项工作定义和探索“3R”-一种新的方法来设计和实现分布式系统，重点是通过战略性地重用环境的视图和回收以前收集的测量，最大限度地减少聚合控制和管理开销。特别是，我们正在设计有效的技术来维护，访问和重用这些信息，以构建下一代流媒体多播，内容分发和数据共享应用程序。这项工作探讨了环境测量的回收/重用的权衡，以及它们在分布式系统设计中的含义。最终，它将促进广域应用的部署，造福整个社会。研究议程是由一个全面的教育和推广计划，以加强实验系统的教育和促进少数招聘和保留的补充。