Ensuring the confidentiality and privacy of sensitive information in cloud computing and outsourced databases is crucial. Homomorphic encryption (HE) offers a solution by enabling computations on encrypted data without decryption, allowing secure outsourcing while maintaining data confidentiality. However, HE faces performance challenges in query-intensive databases. To address this, we propose two novel optimizations, Silca and SilcaZ, tailored to outsourced databases in cloud computing. Silca utilizes a singular caching technique to reduce computational overhead, while SilcaZ leverages modular arithmetic operations to ensure the applicability of singular caching for intensive HE operations. We prove the semantic security of Silca and SilcaZ and implement them with CKKS and BGV in HElib as MySQL loadable functions. Extensive experiments with seven real-world datasets demonstrate their superior performance compared to existing HE schemes, bridging the gap between theoretical advancements and practical applications in applying HE schemes on outsourced databases in cloud computing.

翻译：摘要：确保云计算和外包数据库中敏感信息的机密性与隐私至关重要。同态加密（HE）提供了一种解决方案，允许在不解密的情况下对加密数据进行计算，从而在保持数据机密性的同时实现安全外包。然而，在查询密集型数据库中，同态加密面临性能挑战。为解决此问题，我们针对云计算中的外包数据库提出了两种新颖的优化方案：Silca和SilcaZ。Silca采用单一缓存技术以降低计算开销，而SilcaZ利用模算术运算确保单一缓存技术适用于密集的同态加密操作。我们证明了Silca与SilcaZ的语义安全性，并在HElib中基于CKKS和BGV方案将其实现为MySQL可加载函数。基于七个真实数据集的广泛实验表明，与现有同态加密方案相比，所提方法具有优越性能，从而弥合了同态加密方案在云计算外包数据库中从理论进展到实际应用之间的鸿沟。