云环境下基于SLA的信任协商机制研究
发布时间：2018-11-23 14:18
【摘要】：随着云计算技术的蓬勃发展,出现了各种参差不齐、质量不一的云服务。由于目前缺少细分的行业服务标准,用户在使用云服务过程中,很可能会遇到服务中断、服务水平过低、甚至是用户数据安全受到威胁等问题,用户的恶意评价对云服务信誉会造成严重影响。因此,实现用户与服务之间的可信协商已经成为服务计算领域的热点关注问题,基于SLA(Service Level Agreement)的协商模式在实体之间可信交互中正发挥着越来越大的作用。本文围绕如何实现用户实体与服务实体之间可信协商交互进行研究,通过数据规范化,用户虚假信息过滤,不负责任评价过滤等多个层面对用户的QoS评价信息进行了处理;在服务信誉值计算中,对直接信任度计算中的权重分配,用户评价可信度,稳定性,影响因素分析等几个方面进行了度量,使直接信任计算更加的准确;在间接信任计算中,通过可信用户推荐,寻找可信路径等方式实现准确的间接可信度计算,解决了用户实体与服务实体之间的间接可信度准确计算问题。本文的主要工作及创新点如下:1.在服务实体信誉值计算方面,提出了一种面向个性化服务属性的间接可信度计算方法,通过分析交互发生时间、使用服务的频率、服务交易金额等诸多因素,精确计算直接可信度。在此基础上,利用树状拓扑结构寻找可信路径的方法,解决了间接可信度准确计算问题。2.在服务等级交互方面,提出了一种云环境下基于SLA的动态交互方法,通过分析判断云服务商的可信级别以及用户的可信级别,拒绝高危险系数的用户访问服务,降低用户恶意访问次数,防止恶意评价数量过多,并且避免用户对不可信服务的重复使用;根据用户可信级别,确定云用户隐私数据的保密级别,从而规范云用户的可信评价行为,对云用户的可信评价做出行为奖惩措施,有效提高服务信誉值计算的精确度,为用户与云服务之间的交互提供可信环境。3.在用户个性化选择服务方面,对用户评价权重进行了分析,通过为QoS各项属性分配不同的权重,体现用户的个性偏好,这直接影响到最终的服务信誉度的计算值。仅对用户比较感兴趣的几个类别进行权重分配计算,既节省了计算资源,又达到了计算结果准确且体现用户个性化的目标。4.在稳定性服务推荐方面,提出了一种基于稳定性与用户可信评价的服务推荐方法,通过将服务稳定性量化的方式,为用户解决服务选择的问题,并且对恶意用户评价进行过滤,减少恶意评价对服务可信度计算的影响。在间接推荐树的基础上,深层次拓展提出Top-k服务推荐网络的概念,精确计算服务推荐值,提高服务信誉值计算的准确度。
[Abstract]:With the rapid development of cloud computing technology, there are many different cloud services with different quality. Due to the lack of service standards in the industry at present, users may encounter some problems such as service interruption, low service level and even the threat to user data security in the process of using cloud services. The malicious evaluation of users will have a serious impact on the reputation of cloud services. Therefore, the implementation of trusted negotiation between users and services has become a hot issue in the field of service computing. Negotiation mode based on SLA (Service Level Agreement) is playing a more and more important role in trusted interaction between entities. This paper focuses on how to realize the trusted negotiation interaction between user entities and service entities. Through data standardization, user false information filtering, irresponsible evaluation filtering and so on, the user's QoS evaluation information is processed. In the calculation of service reputation value, the weight distribution, user evaluation reliability, stability and influence factor analysis of direct trust calculation are measured to make the direct trust calculation more accurate. In indirect trust computing, accurate indirect trust calculation is realized by recommending trusted users and finding trusted paths, which solves the problem of accurate calculation of indirect credibility between user entities and service entities. The main work and innovation of this paper are as follows: 1. In the aspect of calculating the credit value of service entity, an indirect credibility calculation method for individualized service attribute is proposed. By analyzing the interaction time, the frequency of using service, the transaction amount of service, and so on, Accurate calculation of direct credibility. On this basis, using the tree topology to find the trusted path, the problem of accurate calculation of indirect credibility is solved. 2. In the aspect of service level interaction, a dynamic interaction method based on SLA in cloud environment is proposed. By analyzing and judging the trust level of cloud service provider and user's trust level, the users with high risk coefficient are denied access to the service. Reducing the number of malicious visits of users, preventing the number of malicious evaluation from excessive, and avoiding the repeated use of untrusted services by users; According to the level of user trust, the level of confidentiality of cloud users' privacy data is determined, so as to standardize the behavior of cloud users' trusted evaluation, to reward and punish the behavior of cloud users' trusted evaluation, and to effectively improve the accuracy of service reputation value calculation. Provide a trusted environment for interaction between users and cloud services. 3. In the aspect of personalized service selection, the weight of user evaluation is analyzed. By assigning different weights to each attribute of QoS, the user's personality preference is reflected, which directly affects the calculation value of the final service reputation. Only several classes of users are interested in weight distribution calculation, which not only saves computing resources, but also achieves the goal of accurate calculation results and personalization of users. 4. In the aspect of stable service recommendation, a service recommendation method based on stability and user trust evaluation is proposed. By quantifying the service stability, the problem of service selection is solved for the user, and the malicious user evaluation is filtered. Reduce the effect of malicious evaluation on service reliability calculation. On the basis of indirect recommendation tree, the concept of Top-k service recommendation network is proposed to accurately calculate service recommendation value and improve the accuracy of service reputation value calculation.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09
本文编号：2351832
[Abstract]:With the rapid development of cloud computing technology, there are many different cloud services with different quality. Due to the lack of service standards in the industry at present, users may encounter some problems such as service interruption, low service level and even the threat to user data security in the process of using cloud services. The malicious evaluation of users will have a serious impact on the reputation of cloud services. Therefore, the implementation of trusted negotiation between users and services has become a hot issue in the field of service computing. Negotiation mode based on SLA (Service Level Agreement) is playing a more and more important role in trusted interaction between entities. This paper focuses on how to realize the trusted negotiation interaction between user entities and service entities. Through data standardization, user false information filtering, irresponsible evaluation filtering and so on, the user's QoS evaluation information is processed. In the calculation of service reputation value, the weight distribution, user evaluation reliability, stability and influence factor analysis of direct trust calculation are measured to make the direct trust calculation more accurate. In indirect trust computing, accurate indirect trust calculation is realized by recommending trusted users and finding trusted paths, which solves the problem of accurate calculation of indirect credibility between user entities and service entities. The main work and innovation of this paper are as follows: 1. In the aspect of calculating the credit value of service entity, an indirect credibility calculation method for individualized service attribute is proposed. By analyzing the interaction time, the frequency of using service, the transaction amount of service, and so on, Accurate calculation of direct credibility. On this basis, using the tree topology to find the trusted path, the problem of accurate calculation of indirect credibility is solved. 2. In the aspect of service level interaction, a dynamic interaction method based on SLA in cloud environment is proposed. By analyzing and judging the trust level of cloud service provider and user's trust level, the users with high risk coefficient are denied access to the service. Reducing the number of malicious visits of users, preventing the number of malicious evaluation from excessive, and avoiding the repeated use of untrusted services by users; According to the level of user trust, the level of confidentiality of cloud users' privacy data is determined, so as to standardize the behavior of cloud users' trusted evaluation, to reward and punish the behavior of cloud users' trusted evaluation, and to effectively improve the accuracy of service reputation value calculation. Provide a trusted environment for interaction between users and cloud services. 3. In the aspect of personalized service selection, the weight of user evaluation is analyzed. By assigning different weights to each attribute of QoS, the user's personality preference is reflected, which directly affects the calculation value of the final service reputation. Only several classes of users are interested in weight distribution calculation, which not only saves computing resources, but also achieves the goal of accurate calculation results and personalization of users. 4. In the aspect of stable service recommendation, a service recommendation method based on stability and user trust evaluation is proposed. By quantifying the service stability, the problem of service selection is solved for the user, and the malicious user evaluation is filtered. Reduce the effect of malicious evaluation on service reliability calculation. On the basis of indirect recommendation tree, the concept of Top-k service recommendation network is proposed to accurately calculate service recommendation value and improve the accuracy of service reputation value calculation.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09
【相似文献】
相关期刊论文 前4条
1 刘迎春;郑小林;陈德人;;信任网络中基于角色信誉的信任预测[J];北京邮电大学学报;2013年01期
2 付晓东;邹平;姜瑛;;基于质量相似度的Web服务信誉度量[J];计算机集成制造系统;2008年03期
3 吴吉义;沈千里;陈德人;章剑林;;融合QoS的对等电子服务信誉激励研究进展[J];华中科技大学学报(自然科学版);2012年S1期
4 付晓东;岳昆;邹平;姜瑛;;随机服务质量感知的Web服务信誉度量模型[J];计算机集成制造系统;2011年08期
相关会议论文 前1条
1 ;江西明骏(香港)实业有限公司董事长 赖达雄[A];时代先锋——中国区域经济杰出人物（第1卷）[C];2006年
相关重要报纸文章 前1条
1 姜云飞;营造软件和信息服务外包国际公信力[N];大连日报;2008年
相关硕士学位论文 前3条
1 樊悦芹;云环境下基于SLA的信任协商机制研究[D];山东师范大学;2017年
2 杨光;基于客观自动反馈的服务信誉评估模型[D];上海交通大学;2008年
3 牟宇迪;基于SLA的端到端Web服务信誉评价方法[D];东北大学;2009年
，本文编号：2351832
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2351832.html
