Onyx: Cost-Efficient Disk-Oblivious ANN Search

📰 ArXiv cs.AI

Learn how Onyx enables cost-efficient disk-oblivious ANN search, protecting sensitive data in AI systems

advanced Published 23 Apr 2026

Action Steps

Implement Onyx to hide disk access patterns in ANN search
Use Trusted Execution Environments (TEEs) to protect sensitive data
Configure Oblivious RAM (ORAM) for secure data access
Test Onyx with existing disk-based ANN search techniques
Compare the cost efficiency of Onyx with other secure ANN search methods

Who Needs to Know This

AI engineers and researchers working with sensitive data on third-party infrastructure can benefit from Onyx's cost-efficient approach to protecting user queries

Key Insight

💡 Onyx protects sensitive data in AI systems by hiding disk access patterns, making it a cost-efficient solution for secure ANN search

Key Takeaways

Learn how Onyx enables cost-efficient disk-oblivious ANN search, protecting sensitive data in AI systems

Full Article

Title: Onyx: Cost-Efficient Disk-Oblivious ANN Search

Abstract:
arXiv:2604.20401v1 Announce Type: cross Abstract: Approximate nearest neighbor (ANN) search in AI systems increasingly handles sensitive data on third-party infrastructure. Trusted execution environments (TEEs) offer protection, but cost-efficient deployments must rely on external SSDs, which leaks user queries through disk access patterns to the host. Oblivious RAM (ORAM) can hide these access patterns but at a high cost; when paired with existing disk-based ANN search techniques, it makes poor

Read full paper → ← Back to Reads