Oblivious RAM for MPC
summary
Claims a practical speedup: secure computation of ORAM using Tree ORAM plus honest‑majority three‑party MPC (one corrupt) replaces costly 2PC garbled circuits. Core tools are Yao‑style garbling, OT extension and a three‑party precomputed OT trick, plus protocol tweaks for eviction. Result shrinks per‑access circuit size to about 69K–233K gates (versus millions previously), keeps asymptotic overhead ≈ O(log^3 N), and reports an implementation that closely matches native ORAM performance.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:28Outline
- 3:48Secure Computation Primitive
- 6:50Application
- 10:29Security
- 12:02Generic
- 16:49Three Parties
- 25:19Technical Details
- 28:04The Client
- 32:04The Eviction
- 33:04Conclusion
related talks
01h 15mOblivious RAM I
01h 06mOblivious RAM II
01h 08mSecure Computation 1
01h 14mPrivate Information Retrieval and Oblivious RAM
29mTowards Scalable Constant-Round MPC from Minimal Assumptions via Round Collapsing
01h 11mPrivate Information Retrieval and Oblivious RAM
40mApplied MPC
01h 03mSecure Computation 2
