基于服务认知和服务评估模型的分组丢弃算法的研究与实现

发布时间：2018-03-08 12:38

本文选题：认知网络　切入点：服务评估模型　出处：《北京邮电大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着网络技术与网络应用规模的快速发展,当前网络上运行的业务种类与业务数量也相对众多,网络环境也变得更加复杂,网络态势更加难以控制,这种情况对传统网络提出了新的要求。在这种情况下,为了保障端对端的服务质量(QoS),认知网络技术应运而生,前瞻性地提出了根据对网络态势的实时感知而动态地对网络配置进行调整的方法,为保障端对端QOS提供了新的思路,也成为当前网络技术范畴的研究热点之一。本文基于认知网络的基本思想,提出了一种基于服务认知和服务评估模型的分组丢弃算法,应用场景为网络发生拥塞的特殊状态。算法模型包括两方面内容：一方面该算法模型提供了底层数据包控制原理的高层服务依据,另一方面算法模型结合服务优先级与服务运行特征两方面内容建立了服务评估模型,在成功地兼顾了网络服务的公平性与效率性的同时,在一定程度上保证了高级别服务的正常运行。主要工作有如下几个方面： (1)提出了基于服务认知和服务评估模型的分组丢弃算法的整体架构,借鉴了传统的分组丢弃算法的核心思想,加入了服务认知功能与服务评估模型用于生成分组丢弃概率参数,从服务层面体现了认知网络的认知思想； (2)建立了有效的服务评估模型,综合多种业务类型与服务运行状态特征等因素,基于历史服务流量统计与服务传输层传输特点分类处理的服务评估方法,体现了服务调节的反馈性,更适合实际网络条件下的应用场景,能够充分优化网络传输性能,从数据层面保障了端对端的QoS。从仿真和系统实现两个角度对基于服务认知和服务评估模型的分组丢弃算法进行了验证,仿真表明,算法能够以较低的控制成本实现对网络高级服务QoS的可靠性保障,反馈机制的引入也在一定程度上体现了公平性原则,从而有效避免了高级服务对网络资源的过分消费,从用户层面提高了网络资源的利用效率。
[Abstract]:With the rapid development of network technology and network application scale, the types and number of services running on the network are relatively numerous, the network environment is becoming more complex, and the network situation is more difficult to control. This situation puts forward new requirements for traditional networks. In this case, in order to ensure the quality of service from end to end, cognitive network technology emerges as the times require. The method of dynamically adjusting the network configuration according to the real-time perception of network situation is put forward prospectively, which provides a new way to guarantee end-to-end QOS, and also becomes one of the research hotspots in the field of network technology. Based on the basic idea of cognitive network, a packet dropping algorithm based on service cognition and service evaluation model is proposed in this paper. The algorithm model includes two aspects: on the one hand, the algorithm model provides the high-level service basis of the underlying packet control principle. On the other hand, the algorithm model combines service priority and service operation characteristics to establish a service evaluation model, which successfully takes into account the fairness and efficiency of network services at the same time. To a certain extent, it ensures the normal operation of high-level services. The main tasks are as follows:. (1) the whole architecture of packet drop algorithm based on service cognition and service evaluation model is proposed. The core idea of traditional packet dropping algorithm is used for reference, and service cognitive function and service evaluation model are added to generate packet drop probability parameters. It embodies the cognitive thought of cognitive network from the service level. (2) an effective service evaluation model is established, which integrates various service types and service running state characteristics, and provides a service evaluation method based on historical service traffic statistics and service transport layer transport characteristics classification. It reflects the feedback of service regulation, is more suitable for the application scenario under the actual network condition, can optimize the network transmission performance fully, and guarantees the end-to-end QoS from the data level. The packet dropping algorithm based on service cognition and service evaluation model is verified from two aspects of simulation and system implementation. The simulation results show that the algorithm can guarantee the reliability of network advanced service QoS with lower control cost. The introduction of feedback mechanism also reflects the principle of fairness to a certain extent, which effectively avoids the excessive consumption of network resources by advanced services, and improves the efficiency of using network resources at the user level.
【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】：TP393.09

【参考文献】