自组织社会网络中的数据管理中间件协议

发布时间：2018-04-24 05:30

  本文选题：自组织社会网络 + 中间件　； 参考：《大连理工大学》2014年博士论文

【摘要】：自组织社会网络为移动用户间提供了基础设施无关的机会性沟通环境。充分利用由现代传感设备采集的用户社会性特征有助于提升和改善移动自组织网络的性能。其中,以社会感知机制为基础设计的中间件对简洁高效地开发创新性的应用与服务具有重要的作用。然而,由于受到多种因素的限制,如缺乏集中管理、设备异构性、无线通信的不可靠性、移动性、资源约束、需要支持多种流量类型等,一些新的挑战也应运而生。其中一个主要挑战是由应用程序和底层协议之间的差异而造成的性能下降和实施效率低下的问题。此外,数据管理也为人类和无线网络系统提出了严峻的挑战,由于目前仅仅忽略或者只在一定程度上考虑了用户社会和移动信息。自组织社会网络可以广泛地应用在会议场所和一些紧急情况下,其实际场景能清晰地说明在合适的无线网络操作中将重要的服务或模块(如数据管理)和中间件设计相结合的重要性。 在目前的工作中,自组织网络框架已经被提出,但从未在数据管理中考虑社会感知问题。为解决这一不足之处,本文专注于为自组织社会网络中间件开发一个正式的框架,在其上可以设计和提出数据管理协议。本文充分利用社会网络和移动自组织网络去整合社会感知和用户移动性。这促使本文探讨了数据管理中间件的各方面问题,包括数据可用性、负载分配和社群中用户合作问题。自组织社会网络不同于社会网络在于其网络结构是经常性断连的。传统上,理想方法被广泛的用于解决数据可用性,甚至是负载分配和用户合作问题。然而,无线链路的质量受多种因素影响,例如移动性、开销和用户自私行为。为应对上述挑战,数据拷贝方法和社群划分概念在本文的研究过程中得到了广泛地应用。据我们所知,这是第一次在自组织社会网络中同时研究数据可用性、负载分配和自私问题。 自组织社会网络服务的可访问性和可用性可由拷贝方法来保证,该方法有助于避免因组群的不确定性移动引起的社群划分以及数据丢失,同时也利于减少数据从源端到目的端经历的跳数。但是由于资源限制,不可能在每个节点上拷贝所有的数据项。问题由此产生,在何处、如何分配数据副本以获得高数据可访问性和可用性。本文首先在该方向上以最大化数据可用性为目标设计拷贝协议ComPAS,一个基于社群划分感知的拷贝分配方法。ComPAS运用一致有效的方式为每位用户存储数据副本,它基于系统的拷贝预算及其所需可用性来选择副本数量。该方法的目标包括利用社会关系为社群中的数据设置拷贝,通过尽可能降低拷贝读取消耗、重定位消耗和通信量,提高有效性和一致性。该方法同时将平衡社群存储空间负载作为另一个目标。与其他方案相比,ComPAS在读取消耗上提供了更引入关注的模式,并且能高效处理网络中的副本重定位问题。 自组织社会网络系统同时面临着性能退化和稳定性差的挑战,这是由工作时的负载不均衡分布和链路失效造成的。分布式系统的正常功能是由其各个计算模块协同工作而实现的。一个有效的负载均衡策略能保证系统资源的优化利用,并且各个代理中不会出现空载或过载的情况。在现如今的许多分布式环境中,包括自组织社会网络,用户由有限的存储空间、带宽等资源相连接从而造成时延,这使得用户需要依靠遍及整个网络的资源通信和负载交换。为了充分利用这类网络系统,数据可用性和资源分发是关键问题,其中负载均衡、分区和容错都是需要解决的问题。因此,本文提出了基于社群的、具有负载均衡与公平机制的事件分发方法Co-Lab,强化了数据管理性能。该方法采用兴趣相似性和过滤拷贝方法在社群内进行代理分簇。Co-Lab的目标是在社群代理间获得更加均匀的负载分布,在保持尽可能高的可靠性下避免过载代理的出现。性能评价部分表明Co-Lab具有明显的优势,可以达到相对更好的负载均衡,降低总体负载,以及对于失效情况的鲁棒性。 现有的数据管理协议的另一不足之处在于进行转发数据等操作时通常会假设用户是合作的。如果节点拒绝合作,端到端的连接就无法形成。这样的非自愿(自私)的行为会严重降低网络的性能。所以,探测自私用户、减轻其对合作用户的影响是必不可少的。尽管在探测自私用户方面已有研究,但是围绕该问题仍存在一些问题,尤其是面对自组织网络中的数据分发和转发服务时。所以,本文意欲设计一种算法,让每位用户能自主地识别和排除自私用户。这类特性可以从细菌等的生物过程中观察得出,并且启发我们将这一新的合作框架的概念用于拷贝分配协议上。因此,本文提出了一种以ComPAS为数据拷贝模型的生物启发式的算法,称作BoDMaS,以探测和减轻自私节点的影响。BoDMaS采用社会和生物机制,对用户进行评估和分类,并拒绝自私用户的参与。为评价该策略的有效性,本文采用了不同的指标。结果表明,在自组织社会网络环境下,BoDMaS能准确探测在拷贝操作中的自私性。
[Abstract]:The self-organized social network provides an infrastructure independent opportunistic communication environment for mobile users. Making full use of the social characteristics of users collected by modern sensing devices helps to improve and improve the performance of mobile ad hoc networks. Among them, middleware based on social perception mechanism is innovative and efficient in developing and innovating. However, some new challenges, such as lack of centralized management, isomeric equipment, unreliability of wireless communications, mobility, resource constraints, need to support a variety of traffic types, and a number of new challenges. One of the major challenges is the application and the underlying protocol. In addition, data management also poses a severe challenge to the human and wireless network systems. Since only a certain degree of neglect or only a certain degree of consideration of the user's social and mobile information, the self organized social network can be widely used in conference sites and some of them. In an emergency, its actual scene clearly illustrates the importance of combining important services or modules (such as data management) and middleware design in a suitable wireless network operation.
In the present work, the self-organizing network framework has been proposed, but the social perception problem has never been considered in the data management. In order to solve this problem, this paper focuses on the development of a formal framework for self-organizing social network middleware, and can design and propose data management protocols on it. This paper makes full use of social networks and Mobile ad hoc networks integrate social perception and user mobility. This has prompted this paper to discuss all aspects of data management middleware, including data availability, load distribution and user cooperation in the community. The self organized social network is different from the social network because its network structure is often disconnected. Traditionally, the ideal method It is widely used to solve data availability, even load distribution and user cooperation. However, the quality of wireless links is affected by many factors, such as mobility, overhead and user selfish behavior. In response to the above challenges, the data copy method and community division concept have been widely used in the research process of this article. For the first time, it is the first time to study data availability, load distribution and selfishness in a self organizing social network.
The accessibility and availability of self-organized social network services can be guaranteed by copy methods, which can help avoid community division and data loss caused by group uncertainty movement, and also reduce the number of jumps from source to destination. However, it is impossible to copy each node by resource constraints. All data items. The problem arises, where, how to assign a copy of data to obtain high data accessibility and availability. First, in this direction, a copy protocol ComPAS is designed to maximize data availability, and a community partition aware copy allocation method.ComPAS uses a consistent and effective way for each A user stores a copy of the data, which is based on the copy budget of the system and its availability to select the number of copies. The aim of this method is to use social relations to copy the data in the community, to reduce the copy reading consumption as much as possible, to relocate the consumption and traffic, to raise the efficiency and consistency. The group storage space load is another goal. Compared with other solutions, ComPAS provides more interesting patterns for reading consumption and can efficiently handle replica relocation problems in the network.
The self organized social network system is confronted with the challenge of performance degradation and poor stability, which is caused by the unbalanced load distribution and link failure at work. The normal function of the distributed system is implemented by the cooperative work of various computing modules. An effective load balancing strategy can ensure the optimal utilization of the system resources. In many distributed environments, including self organized social networks, users are connected by limited storage space, bandwidth and other resources to cause delay, which makes users rely on resource communication and load exchange throughout the network. In order to make full use of this Network system, data availability and resource distribution are the key problems, in which load balancing, partition and fault tolerance are all problems to be solved. Therefore, this paper proposes an event distribution method, Co-Lab, based on community, with load balancing and fairness mechanism, which strengthens the performance of data management. This method uses interest similarity and filter copying. The objective of the proxy clustering.Co-Lab in the community is to obtain a more uniform load distribution between the community agents and avoid the appearance of the overloading agent under the highest possible reliability. The performance evaluation part shows that Co-Lab has obvious advantages to achieve a relatively better negative load balance, the overall load reduction, and the failure of the failure. The robustness of the case.
Another disadvantage of existing data management protocols is that a user is usually assumed to be cooperative when the data is forwarded. If the node refuses to cooperate, the end to end connection can not be formed. Such an involuntary (selfish) behavior will seriously reduce the performance of the network. Therefore, it will detect selfish users and reduce their shadow to the cooperative users. Noise is essential. Although there has been a study on the detection of selfish users, there are still some problems around this problem, especially in the face of data distribution and forwarding services in a self organized network. So, this article intends to design an algorithm that allows each user to recognize and exclude selfish users independently. This kind of feature can be from bacteria. The concept of this new cooperative framework is used to apply the concept of this new cooperative framework to the copy allocation protocol. Therefore, this paper proposes a bioheuristic algorithm based on ComPAS as a data copy model, called BoDMaS, to detect and mitigate the influence of selfish nodes on the social and biological mechanisms, and to the users. In order to evaluate the effectiveness of the strategy, different indicators are used to evaluate the effectiveness of the strategy. The results show that BoDMaS can accurately detect the selfishness in the copy operation under the self-organized social network environment.

【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.5
，

本文编号：1795369