• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Design of availability-dependent distributed services in large-scale uncooperative settings.
【24h】

Design of availability-dependent distributed services in large-scale uncooperative settings.

机译：在大规模不合作的环境中设计依赖于可用性的分布式服务。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Thesis Statement: Availability-dependent global predicates can be efficiently and scalably realized for a class of distributed services, in spite of specific selfish and colluding behaviors, using local and decentralized protocols.;Several types of large-scale distributed systems spanning the Internet have to deal with availability variations among their constituent nodes. In dealing with churn and low availability nodes, we believe it is important to link the availability of a node to the service the node receives from the distributed system. In other words, high availability has to be incentivized with better service. There are two types of requirements for this problem. First, metrics such as message overhead, CPU usage, memory overhead and latency need to be optimized to achieve scalability and efficiency. Secondly, in open distributed systems spanning multiple organizations, the protocols have to tolerate selfish and colluding nodes, i.e., low availability nodes that attempt to receive better service.;This thesis approaches this problem by explicitly linking each node's service to its availability, via the notion of a global predicate. We present a class of novel distributed protocols that achieve a given availability-dependent global predicate, efficiently and scalably. These protocols execute in a fully decentralized manner, realizing the global predicates in an emergent fashion. Predicate satisfaction is resilient to churn, and to selfish and colluding nodes. The eventual goal of the predicates is to help incentivize nodes to improve their availability in order to get better service.;Our approach includes using random and consistent techniques to build overlays, as well as probabilistic local actions such as message forwarding, monitoring, and auditing. This combination of techniques leads to realizing the predicates, and to probabilistic tolerance to failures, both churn-related as well as from selfish and colluding behaviors.;Concretely, this thesis makes three major contributions that are closely related to each other. First we present AVMON, the first distributed availability monitoring service. AVMON builds random and consistent overlays for accurate and decentralized monitoring of the long term availability of each node. Second, we present AVMEM, the first availability-aware overlay. Nodes in AVMEM build their membership by using a globally assigned predicate, leveraging our AVMON work. On top of AVMEM we implement management functions that query nodes based on their availability---range/threshold multicast and range/threshold anycast. Finally, we present AVCOL, the first availability-aware network aggregation system that realizes availability-dependent predicates. The predicates specify the probability that a node's aggregate becomes part of the global aggregate, as a function of the node's availability.;We evaluate our systems through mathematical analysis, and thorough experimentation. We carry out our experiments using synthetic and real system traces. Our results demonstrate the probabilistic correctness, scalability, fault-tolerance, and efficiency of our protocols.
机译：论文陈述：尽管存在特定的自私和共谋行为，但使用本地和分散协议，可以针对一类分布式服务有效且可扩展地实现与可用性相关的全局谓词。跨越Internet的几种大型分布式系统必须处理其组成节点之间的可用性差异。在处理客户流失和低可用性节点时，我们认为将节点的可用性与该节点从分布式系统接收的服务联系起来很重要。换句话说，必须通过更好的服务来激励高可用性。此问题有两种类型的要求。首先，需要优化诸如消息开销，CPU使用率，内存开销和延迟之类的指标，以实现可伸缩性和效率。其次，在跨越多个组织的开放式分布式系统中，协议必须容忍自私的和共谋的节点，即试图获得更好服务的低可用性节点。本文通过将每个节点的服务通过其显式链接到其可用性来解决此问题。全局谓词的概念。我们提出了一种新颖的分布式协议，可以高效，可扩展地实现给定的依赖于可用性的全局谓词。这些协议以完全分散的方式执行，以一种新兴的方式实现全局谓词。谓词满足对于流失，自私和共谋节点具有弹性。谓词的最终目标是帮助激励节点以提高其可用性以获得更好的服务。；我们的方法包括使用随机且一致的技术来构建覆盖图，以及使用概率性本地操作（例如消息转发，监视和审核） 。这种技术的结合导致谓词的实现，以及对失败的概率容忍度，无论是与搅动有关的还是与自私和共谋行为有关的失败。；具体地说，本论文做出了三个密切相关的主要贡献。首先，我们介绍AVMON，这是第一个分布式可用性监视服务。 AVMON构建随机且一致的覆盖图，以对每个节点的长期可用性进行准确和分散的监控。其次，我们介绍AVMEM，这是第一个可用性感知覆盖图。 AVMEM中的节点利用我们的AVMON工作，通过使用全局分配的谓词来建立其成员资格。在AVMEM之上，我们实现了管理功能，这些功能可根据节点的可用性来查询节点-范围/阈值多播和范围/阈值任意播。最后，我们介绍AVCOL，这是第一个实现可用性相关谓词的可用性感知网络聚合系统。谓词根据节点的可用性指定节点的聚合成为全局聚合的一部分的可能性。我们通过数学分析和全面的实验评估我们的系统。我们使用合成的和真实的系统轨迹进行实验。我们的结果证明了我们协议的概率正确性，可伸缩性，容错性和效率。

著录项

  • 作者

    Morales, Ramses Victor.;

  • 作者单位

    University of Illinois at Urbana-Champaign.;

  • 授予单位 University of Illinois at Urbana-Champaign.;
  • 学科 Computer Science.
  • 学位 Ph.D.
  • 年度 2009
  • 页码 105 p.
  • 总页数 105
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号