Pindex: Private multi-linked index for encrypted document retrieval.

Area of Science:

• Computer Science
• Cryptography
• Data Security

Background:

• Cloud storage is essential for outsourcing sensitive data, but updates pose confidentiality risks due to potential information leakage.
• Existing dynamic searchable encryption schemes often struggle to provide forward privacy guarantees against untrusted servers.
• File-injection attacks highlight the need for secure update mechanisms in encrypted cloud storage.

Purpose of the Study:

• To design an efficient searchable encryption scheme that supports dynamic updates while guaranteeing forward privacy.
• To address the challenge of preventing information leakage about past queries during data updates in cloud storage.
• To introduce a novel dynamic index structure for secure and private encrypted document retrieval.

Main Methods:

• Proposed a novel private multi-linked dynamic index named Pindex for encrypted document retrieval.
• Constructed the index using a probabilistic homomorphic encryption mechanism and secret orthogonal vectors.
• Provided rigorous security proofs for correctness and forward privacy within the random oracle model.

Main Results:

• Pindex demonstrates practical efficiency and effectiveness on the real-world Enron dataset.
• The proposed dynamic multi-linked index successfully guarantees forward privacy.
• Achieved forward privacy without a significant compromise in the scheme's overall efficiency.

Conclusions:

• Pindex offers a practical and efficient solution for secure dynamic updates in cryptographic cloud storage.
• The scheme effectively ensures forward privacy, protecting user confidentiality against untrusted servers.
• Pindex represents a significant advancement in searchable encryption for sensitive data management in the cloud.