综合传感网业务数据管理技术研究

发布时间：2018-12-16 00:25

【摘要】：综合传感网(ISN)切合下一代网络发展,将通信网络和传感网络有机融合,为互联网的发展与应用提供了可行的方案。ISN中业务数据呈现分散化、多样化、多速化特点,必须建立高效数据存取与处理平台,实现ISN不同层次结构中数据的科学管理。本文提出开展ISN数据管理技术研究,从管理架构、建模分析和具体实现三个方面开展深入研究,有着重要的意义。论文首先介绍了ISN业务数据管理功能框架,设计了业务数据管理系统分层方案,包括数据处理层、存储层和用户查询层,提出了衡量业务数据管理系统性能的技术指标。接着,提出并建立了单节点业务数据管理模型,提出了基于M/G1/1多层次多优先级排队算法,仿真结果表明,随着到达速率增加,各优先级延时增加,较高优先级的延时增长速率较缓慢,而较低的优先级的延时增长速率较快,到达速率和流量的比值增加。提出并建立了多节点间业务数据管理模型,提出了动态负载均衡可调一致性哈希算法,讨论了静态初始化、动态数据分配等方法。建立了多队列排队数据管理系统,基于动态负载均衡可调一致性哈希算法和普通一致性哈希算法,分别分析了系统负载均衡率。结果表明,前者平均负载率方差为5.37×10-6,后者为8.67×10-5,采用前者后系统负载均衡方差减小了一个数量级,负载均衡性能明显提升。最后,建立了基于以太网的10个节点分布式ISN业务数据管理系统,分析了系统的负载均衡率,实现了业务数据实时监控、实时、定时同步和查询功能。实验结果表明,系统中各数据库负载均衡率均小于0.10,动态负载均衡可调一致性哈希算法可保证系统负载均衡性能。
[Abstract]:The integrated sensor network (ISN) is suitable for the development of the next generation network and integrates the communication network and the sensor network organically, which provides a feasible scheme for the development and application of the Internet. The service data in ISN is distributed, diversified and multi-speed. It is necessary to establish an efficient data access and processing platform to realize the scientific management of data in different hierarchies of ISN. This paper puts forward the research of ISN data management technology, which is of great significance from three aspects of management architecture, modeling analysis and concrete implementation. Firstly, the paper introduces the functional framework of ISN business data management, designs a hierarchical scheme of business data management system, including data processing layer, storage layer and user query layer, and puts forward the technical index to measure the performance of business data management system. Then, a single node service data management model is proposed and a multi-level and multi-priority queuing algorithm based on M/G1/1 is proposed. The simulation results show that with the increase of arrival rate, the delay of each priority increases. The delay growth rate of higher priority is slower, while that of lower priority is faster, and the ratio of arrival rate to flow is increased. A multi-node service data management model is proposed and a dynamic load balancing adjustable consistency hash algorithm is proposed. The static initialization and dynamic data allocation methods are discussed. A multi-queue queueing data management system is established. The system load balancing rate is analyzed based on dynamic load balancing adjustable consistency hash algorithm and common consistent hash algorithm. The results show that the average load rate variance of the former is 5.37 脳 10 ~ (-6) and the latter is 8.67 脳 10 ~ (-5). The system load balancing variance decreases by one order of magnitude and the load balancing performance is improved obviously. Finally, a 10-node distributed ISN service data management system based on Ethernet is established, the load balancing rate of the system is analyzed, and the real-time monitoring, real-time, timing synchronization and query functions of the service data are realized. The experimental results show that the load balancing rate of each database in the system is less than 0.10, and the dynamic load balancing adjustable consistency hashing algorithm can guarantee the load balancing performance of the system.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.52;TP212

【相似文献】