Peeking into the Future: MPC Resilient to Super-Rushing Adversaries
summary
Defines a super-rushing adversary that forces some honest parties into round r+1, reads their round r+1 messages, then crafts remaining round r messages. Shows a concrete break: a 5-party simultaneous-broadcast protocol (one corruption) becomes insecure by peeking one round. Proves a positive general theorem: perfect secret-sharing based protocols (e.g. BGW, multiplication-triplet designs) remain secure unchanged when all-to-all comms hold and the output round is ≥ two rounds after the committed round. Notes statistical security and sequential composition can still fail.
AI-written summary · about
chapters (10)
- 0:00Super-rushing adversaries in MPC
- 2:14Synchronous MPC models
- 10:59Super-rushing adversaries defined
- 13:05Breaking simultaneous broadcast
- 17:48Input providers in MPC security
- 20:20Committed rounds for attack prevention
- 23:38Round gaps to mitigate attacks
- 26:40Positive results for MPC protocols
- 28:29Trade-offs in sequential composition
- 30:37Conclusion on super-rushing resilience
related talks
01h 08mSecure Computation 1
33mInteraction Preserving Compilers for Secure Computation
01h 31mIntroduction to Secure Multi-Party Computation
40mApplied MPC
58mSecure Multiparty Computation I
01h 10mSecure Multiparty Computation II
01h 03mSecure Computation 2
26mEfficient Multiparty Protocols via Log-Depth Threshold Formulae