网络实时分析系统的节点管理功能的设计与实现

发布时间：2018-02-21 19:18

本文关键词： 实时分析分布式对等网络节点管理　出处：《北京邮电大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着网络速度和网络数据流量的不断增长,网络数据分析的实时性需求越来越高。单一主机的处理性能不足促进了分布式实时数据处理系统的快速发展,对此类分布式系统的各个节点进行有效管理成为重要的研究内容。在研究现有的分布式系统节点管理功能的基础上,确定了网络实时分析系统节点管理功能的设计目标：针对系统的“去中心节点的对等架构”的节点集群进行管理,该网络实时分析系统支持搭建可伸缩的节点群集,节点管理功能支持系统的可扩展性,节点管理功能可以处理集群中部分节点发生故障的情况,当某些处理节点发生故障无法正常工作时,由其他节点接手故障节点的工作。本论文首先介绍根据流计算模型开发的分布式系统——网络实时分析系统,重点描述该系统的总体架构与节点管理相关部分,然后分析网络实时分析系统的节点管理的详细需求,包括对数据适配节点和数据分析节点进行状态监控和有效管理,对各个数据分析节点的负载任务进行负载均衡,具体来说,节点管理可以详细分为五个部分：1)节点配置管理功能,2)节点的增删改查基本操作功能,3)节点保活功能,4)节点故障管理功能,5)节点负载均衡功能。之后根据需求设计出详细的相关方案和架构,在实现以上功能之后对网络实时分析系统节点管理各项功能进行全面的功能和性能测试,验证了本文设计实现的网络实时分析系统节点管理在功能和性能方面的优势。最后对本文设计实现的网络实时分析系统节点管理提出建议和下一步研究方向。
[Abstract]:With the increasing of network speed and network data flow, the real-time requirement of network data analysis is more and more high. The lack of processing performance of single host promotes the rapid development of distributed real-time data processing system. The effective management of each node in this kind of distributed system becomes an important research content. On the basis of studying the existing node management function of distributed system, the design goal of node management function of network real-time analysis system is determined: to manage the node cluster of "peer-to-peer architecture of de-central node" in the system. The network real-time analysis system supports the construction of scalable node cluster, the node management function supports the expansibility of the system, and the node management function can deal with the failure of some nodes in the cluster. When some processing nodes fail to work properly, other nodes take over the work of the failed nodes. In this paper, we first introduce the distributed network real-time analysis system, which is developed according to the flow computing model, and mainly describe the overall architecture of the system and the related parts of node management. Then the detailed requirements of node management of network real-time analysis system are analyzed, including monitoring and effective management of data adaptation node and data analysis node, load balancing of each data analysis node, specifically, Node management can be divided into five parts in detail: 1) node configuration management function. Figure out the details of the relevant programs and structures, After realizing the above functions, the functions of node management in the network real-time analysis system are tested. The advantages of node management in the function and performance of the network real-time analysis system designed and implemented in this paper are verified. Finally, the suggestions and the next research direction of node management of the network real-time analysis system designed and implemented in this paper are put forward.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【相似文献】