分布式环境下基于倒排索引的可搜索加密研究

Research on Searchable Encryption Based on Inverted Index in Distributed Environment

摘要: 目的：大数据、云计算技术和分布式技术给人们带来便捷，但是将个人数据加密并上传到云端服务器存储带来了个人隐私泄漏的问题。可搜索加密技术实现了安全高效地检索密文。方法：本文基于Spark计算框架，提出了分布式可搜索加密方案。使用高效的伪随机标签和文档标签通过Spark集群的RDD操作分布式生成倒排索引结构，利用Spark集群的性能优势进行检索，并结合提出的验证算法，进一步提高了分布式可搜索加密方案在半诚实且好奇的威胁模型下的安全性。结果：我们在不同节点数量下考察集群大小对存储性能和对计算效率影响，实验表明集群环境可以有效缓解单机存储的压力，并提升加解密、索引构造和验证的效率。结论：验证了分布式环境下基于倒排索引的可搜索加密的优越性。

Abstract: Purposes: Big data, cloud computing technology and distributed technology bring convenience to people, but encrypting and uploading personal data to the cloud server storage brings about the problem of personal privacy leakage. Searchable encryption technology enables secure and efficient retrieval of ciphertext. Methods: Based on Spark computing framework, this paper proposes a dis-tributed searchable encryption scheme. The efficient pseudo-random tags and document tags are used to generate the inverted index structure through the RDD operation of the Spark cluster, and the performance advantages of the Spark cluster are used for retrieval. Combined with the pro-posed verification algorithm, the security of the distributed searchable encryption scheme under the semi honest and curious threat model is further improved. Finding: We investigate the impact of cluster size on storage performance and computing efficiency under different node numbers. Ex-periments show that the cluster environment can effectively relieve the pressure of stand-alone storage, and improve the efficiency of encryption and decryption, index construction and verifica-tion. Conclusions: The advantages of searchable encryption based on inverted index in distributed environment are verified.