Hardware and software configurations.

The utilization of on-demand remote cloud services provides a flexible way to fulfill the demands of emerging resource-intensive applications. However, migrating data to the cloud also introduced security threats, including unauthorized access and information theft. To resolve this issue, the existing solutions encrypt information locally before uploading it to the server. This process provides information protection with the limitation of non-searchable data. To overcome this limitation, searchable encryption has emerged as a promising cryptographic technique. Some existing searchable encryption techniques are facing data leakage issues by exposing search queries or data to the cloud service provider. Another class of existing searchable schemes introduces processing cost or communication overhead for the data user. The recent searchable solution that is both secure and efficient for data users is Labeled Searchable Encryption (LSE). However, LSE cannot manage large datasets effectively and introduces communication overhead on the data user side. To ensure that Secure Searchable Encryption (SSE) can meet the demands of modern data-driven applications without compromising security and performance, this study aims to investigate and develop novel approaches to enhance the efficiency and security of SSE for large datasets. Experimental findings have proved that the proposed BloomSec is much more efficient and scalable than the classic method of Labeled Searchable Encryption (LSE), consuming significantly less overhead for users, which makes it practically useful for a large dataset without compromising security.

按需远程云服务的应用，为满足新兴资源密集型应用的需求提供了灵活途径。然而，将数据迁移至云端同时也带来了安全威胁，包括未授权访问与信息窃取。为解决该问题，现有解决方案会在将数据上传至服务器前，于本地对信息进行加密。该流程可实现信息保护，但存在数据不可检索的局限性。为克服这一局限，可搜索加密（searchable encryption）已成为一种极具前景的密码学技术。部分现有可搜索加密技术会将搜索查询或数据暴露给云服务提供商，从而面临数据泄露问题。另一类现有可搜索加密方案则会为数据使用者带来处理开销与通信负担。近期出现的、可同时保障数据使用者安全与效率的可搜索加密方案为标记式可搜索加密（Labeled Searchable Encryption, LSE）。但LSE无法有效管理大规模数据集，且会为数据使用者侧带来通信开销。为确保安全可搜索加密（Secure Searchable Encryption, SSE）能够在不降低安全性与性能的前提下，满足现代数据驱动型应用的需求，本研究旨在探索并开发全新方法，以提升面向大规模数据集的SSE的效率与安全性。实验结果证明，本文提出的BloomSec方案相较于经典的标记式可搜索加密（LSE）方案，具备更优异的效率与可扩展性，且可为用户大幅降低开销，因此在不牺牲安全性的前提下，可切实应用于大规模数据集场景。