论文标题

层次缓存的线性函数检索安全性和隐私约束

Hierarchical Cache-Aided Linear Function Retrieval with Security and Privacy Constraints

论文作者

Kong, Yun, Wu, Youlong, Cheng, Minquan

论文摘要

服务器与多个镜像站点连接的分层缓存系统,每个镜像与不同的用户连接,并且镜像站点和用户都配备了缓存记忆。但是,所有现有的作品都集中在单个文件检索上,即每个用户请求一个文件,而忽略通信中的安全性和隐私威胁。在本文中,我们研究了具有内容安全性和需求隐私的层次缓存系统的线性函数检索问题,即,每个用户请求文件的线性组合,同时库中的文件受到窃听程序的保护,并且用户的需求是其他用户未知的,并且用户的需求是其他用户未知的。首先,我们提出了一个名为“层次放置递送阵列(HPDA)”的新组合结构,该结构表征了编码缓存方案的数据放置和交付策略。然后,我们构建了两个类别的HPDA。因此,分别获得有或没有安全性和隐私的两类方案,在第一个专门将第一个跳跃的传输负载最小化,如果忽略安全性和隐私约束,则可以为第一次跳跃实现最佳传输负载;第二个在内存大小上具有更灵活的参数,与第一个相比,子包装较低,并实现了子包装和传输负载之间的权衡。

The hierarchical caching system where a server connects with multiple mirror sites, each connecting with a distinct set of users, and both the mirror sites and users are equipped with caching memories has been widely studied. However all the existing works focus on single file retrieval, i.e., each user requests one file, and ignore the security and privacy threats in communications. In this paper we investigate the linear function retrieval problem for hierarchical caching systems with content security and demand privacy, i.e., each user requests a linear combination of files, and meanwhile the files in the library are protected against wiretappers and users' demands are kept unknown to other users and unconnected mirror sites. First we propose a new combination structure named hierarchical placement delivery array (HPDA), which characterizes the data placement and delivery strategy of a coded caching scheme. Then we construct two classes of HPDAs. Consequently two classes of schemes with or without security and privacy are obtained respectively where the first dedicates to minimizing the transmission load for the first hop and can achieve the optimal transmission load for the first hop if ignoring the security and privacy constraints; the second has more flexible parameters on the memory sizes and a lower subpacketization compared with the first one, and achieves a tradeoff between subpacketization and transmission loads.

扫码加入交流群

加入微信交流群

微信交流群二维码

扫码加入学术交流群,获取更多资源