基于Web的社交物联网体系架构、模型及其安全与隐私关键技术研究

发布时间：2018-08-26 21:05

【摘要】：物联网(Internet of things,IoT)将物理世界的各类传感器、控制器、嵌入式设备等通过有线、无线接入方式连接到互联网中,扩展了信息系统获取物理世界信息的边界和能力,在传统以人类语义为核心的信息网络基础上,引入物体贡献数据的新型信息维度,形成物理世界与人类信息领域融合的抽象数字生态系统。与互联网相比,物联网具有如下特点：(1)物联网部署了海量异构传感器,每个传感器就是一个信息源,不同类型的传感器所获得的信息内容和信息格式不同；(2)物联网终端采集的环境数据具有实时性；(3)以用户体验为中心的应用创新是物联网发展动力。当前以行业应用为主的物联网系统,采用了紧耦合封闭式架构,使系统之间呈现出竖井化的信息孤岛特征,这种架构特征虽然适合需求明确的大规模行业应用,但由于系统基础设施、感知数据和服务能力难以开放和共享,第三方资源也难以被低成本地集成进系统,使得物联网很难大规模的应用和推广。另外,当前物联网领域在业务应用、产业形态方面还存市场碎片化,应用规模小、开发门槛高、开发周期长的、用户参与度低等问题,构成了物联网发展的瓶颈。因此,随时随地实现物联网终端信息数据和控制功能的便捷共享,将感知网络末梢节点融合到信息空间,开放普适服务能力实现泛在计算业务环境,是物联网技术发展的必然方向。针对上述问题,本文将物联网未来发展总结为开放化和社交化两大趋势,提出了基于Web的社交物联网(Social Web of Things, SWoT)体系架构。从开放化角度,SWoT的核心思想为：采用Web服务中先进的面向资源、松耦合的REST架构风格,将物联网中的异构设备、能力抽象成统一的资源模型,以URI (Uniform Resource Identifier)进行统一细粒度资源标识,以屏蔽底层异构性,实现对资源的统一协同处理与协作调度；基于SWoT业务平台和业务网关设计了REST API,以构建更加开放的网络与信息协同业务环境,支持未来物联网中各种服务和资源的开放,以及各种业务应用的快速生成和开发。从社交化角度,SWoT的核心思想是：一方面,支持人与物、物与物之间的社会化智能交互,用户能够共享个人的设备和感知数据；另一方面,将物联网中的实体及其物理、逻辑上的关联关系抽象为社交网络中的点和边,复杂的社交模型可以将“点状”资源通过多维社交关系连接成网,为资源共享、信任度评估、访问控制等服务创造更大空间。本文基于SWoT基本思想,开发了智能家居系统,实现了资源的统一标识、开放,以及节点的智能交互。同时,基于新浪微博社交平台开发了物联网资源发现应用,并进行了性能分析。然而,物联网通过摄像头、传感器等收集的数据信息与人们的日常生活密切相连,并且其感知关联度很大,这个庞大的系统如果不加于安全隐私的规范,将成为一个庞大的盯梢系统。特别是在SWoT物联网业务环境中,通信、数据协同等操作将更加频繁,第三方应用通过开放接口获取数据进行数据挖掘,将会比我们自己更清楚我们的行为、爱好,这将对敏感数据和隐私信息构成威胁；由于感知数据可能存储在异构的分布式业务网关上并进行存储、管理和开放,也会造成存储及搜索安全问题。随着物联网社交化趋势,用户将更多的共享个人设备和感知数据,而多样的SWoT节点属性、复杂的社交关系以及社交过程中的用户生成内容(User Generated Content, UGC)将能够唯一定位一个节点,因此面向传统数据集的安全隐私技术难以直接应用,对用户的隐私泄露更加难以预测,进行隐私保护也更困难。针对物联网社交化带来的安全隐私问题,本文提出了面向攻击者背景知识的数据匿名发布隐私度量算法。首先分析了SWoT中可能存在的攻击者背景知识,针对节点属性相关性和连接关系背景知识,我们设计了基于多变量核回归模型的隐私度量算法,本算法充分考虑各属性的数据分布并提出了局部核带宽的概念,并引入了协方差矩阵对原始假设属性独立的核回归算法进行了改进。针对UGC背景知识,我们根据敏感属性数据类型(连续或分类),分别基于多元线性回归模型和朴素贝叶斯分类模型设计了隐私度量算法。考虑到UGC通过TF-IDF生成的内容向量的维度较高,为减少计算开销,我们利用PCA降维技术对算法进一步进行改进。通过对新浪微博真实数据进行Matlab仿真,验证了本文算法能有效度量由上述三种攻击者背景知识带来的隐私泄露。针对SWoT中由业务网关开放模式带来的存储和搜索安全隐私问题,本文提出了基于节点信任度和P2P Chord虚拟安全域的安全增强搜索算法。我们综合考虑了Underlay物理节点(SWoT业务网关)的异构性和不稳定性,以及Overlay节点在交互过程中的相互评价,设计了Underlay与Overlay融合的信任度评价算法。通过构建基于Chord的虚拟安全域,定义节点域问路由表、域内路由表和评价列表,并讨论了虚拟安全域的维护,以解决SWoT业务网关开放模式下的资源存储、搜索安全问题。同时,讨论了基于P2P虚拟安全域的安全增强搜索算法,设计了面向安全域和节点的两层访问控制机制,以及最优节点选择策略。通过仿真,证明了本算法具有低维护开销和良好的可扩展性,同时能提高搜索过程中的资源访问安全和节点安全。
[Abstract]:Internet of Things (IoT) connects all kinds of sensors, controllers and embedded devices in the physical world to the Internet through wired and wireless access. It expands the boundary and ability of information system to obtain the information of the physical world. On the basis of the traditional information network with human semantics as the core, the object contribution data is introduced. Compared with the Internet, the Internet of Things has the following characteristics: (1) the Internet of Things has deployed a large number of heterogeneous sensors, each sensor is an information source, different types of sensors obtain different information content and information format; (2) Things. The environmental data collected by Internet terminals are real-time; (3) User experience-centered application innovation is the driving force for the development of the Internet of Things.
The current industry-based Internet of Things (IOT) systems adopt a tightly coupled closed architecture, which makes the systems present a silo-like information island. Although this architecture is suitable for large-scale industrial applications with a clear demand, it is difficult to open and share data and service capabilities because of the system infrastructure, and third-party resources. It is difficult to be integrated into the system at low cost, which makes it difficult to apply and popularize the Internet of Things on a large scale. In addition, the problems of fragmentation of market, small scale of application, high threshold of development, long development cycle and low participation of users in the field of Internet of Things constitute the bottleneck of the development of the Internet of Things. It is the inevitable direction of the development of the Internet of Things technology to realize the convenient sharing of information data and control functions of the Internet of Things terminals, to integrate the end nodes of the perceptual network into the information space, and to realize the ubiquitous computing business environment with the open and universal service capability.
In view of the above problems, this paper summarizes the future development of the Internet of Things as two major trends of openness and socialization, and proposes a Web-based Social Web of Things (SWoT) architecture. Heterogeneous devices, capabilities abstracted into a unified resource model, using URI (Uniform Resource Identifier) to unify fine-grained resource identification, to shield the underlying heterogeneity, to achieve unified collaborative processing and scheduling of resources; based on SWOT service platform and service gateway, a REST API is designed to build a more open network and information collaboration. From a social point of view, the core idea of SWoT is: on the one hand, it supports the socialized intelligent interaction between people and things, and between things, so that users can share personal devices and perceived data; on the other hand, it will Entities and their physical and logical relationships in the Internet of Things are abstracted as points and edges in social networks. Complex social models can connect "point-like" resources into a network through multi-dimensional social relations, creating more space for resource sharing, trust evaluation, access control and other services. Home system realizes the unified identification of resources, openness, and intelligent interaction of nodes. At the same time, based on Sina Weibo social platform, the application of Internet of Things resource discovery is developed, and the performance analysis is carried out.
However, the data collected by the Internet of Things (IOT) through cameras, sensors and so on are closely related to people's daily life, and their perceptions are highly related. This huge system will become a huge peer-to-peer system if it is not added to the security and privacy standards. Especially in the SWoT IOT business environment, communication, data collaboration and other operations. More frequently, third-party applications will acquire data through open interfaces for data mining, and will be more aware of our behavior and preferences than we do, which will pose a threat to sensitive data and privacy information; as perceived data may be stored on heterogeneous distributed business gateways and stored, managed, and opened, it will also create storage Storage and search security issues. As the Internet of Things becomes more socialized, users will share more personal devices and perceived data. Various SWOT node attributes, complex social relationships, and user-generated content (UGC) in social processes will be able to uniquely locate a single node, thus providing security for traditional data sets. Privacy technology is difficult to directly apply, and it is more difficult to predict the privacy leakage of users, and it is also more difficult to protect privacy.
Aiming at the security and privacy problems brought about by the socialization of the Internet of Things, this paper proposes a privacy measurement algorithm for anonymous publishing of data based on attacker's background knowledge. Firstly, we analyze the possible attacker's background knowledge in SWoT, and design a hidden algorithm based on multivariate kernel regression model for the correlation of node attributes and the background knowledge of connection relationship. Private measurement algorithm, this algorithm fully considers the data distribution of each attribute and proposes the concept of local kernel bandwidth, and introduces covariance matrix to improve the original hypothesis attribute independent kernel regression algorithm. Considering the high dimensionality of content vectors generated by UGC through TF-IDF, we use PCA dimensionality reduction technology to further improve the algorithm. By simulating the real data of Sina Weibo with MATLAB, we verify that the algorithm can effectively measure the above three. Privacy breaches brought about by background knowledge of attackers.
In order to solve the security and privacy problems of storage and search caused by the open mode of service gateway in SWoT, a security enhanced search algorithm based on node trust and P2P Chord virtual security domain is proposed. A trust evaluation algorithm based on the integration of Underlay and Overlay is designed. The virtual security domain based on Chord is constructed, and the inter-domain routing table, intra-domain routing table and evaluation list are defined. The maintenance of virtual security domain is discussed to solve the problem of resource storage and security search in the open mode of SWoT service gateway. The security enhancement search algorithm based on P2P virtual security domain is discussed. A two-tier access control mechanism for security domain and node is designed, and an optimal node selection strategy is proposed.
【学位授予单位】：北京邮电大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP391.44;TN929.5

【相似文献】