Search papers, labs, and topics across Lattice.
G-Lox introduces a privacy-preserving bridge distribution system that maintains distributor blindness while allowing for adaptive group-level assignments. By employing a two-server architecture with secure two-party computation, G-Lox ensures that no single server has access to group identifiers or assignments, enhancing privacy during state-dependent updates. Evaluations demonstrate that G-Lox achieves low overhead in communication and improved robustness compared to existing systems, making it a significant advancement in privacy-preserving distributed systems.
G-Lox achieves robust, privacy-preserving group adaptations without compromising on performance, outperforming traditional systems in both efficiency and security.
We present G-Lox (group-adaptive Lox), a bridge-distribution system that preserves Lox-style distributor blindness while enabling hidden, stateful group-level adaptation. G-Lox places adaptive assignment logic behind a two-server privacy wall, so no single server learns group identifiers or group-to-bridge assignments. Private state access and state-dependent updates use two-server DPF/FSS protocols and secure two-party computation, supporting blockage reporting, transport-aware reassignment, and privacy-preserving group splitting. We evaluate G-Lox through system measurements and policy simulation. In our C++/EMP implementation over real TCP sockets, private state access has low client-visible overhead: across state sizes up to 2^16, communication remains in the low-KiB range per iteration. At M=1024, the client sends 1,968 bytes, receives 1,280 bytes, and completes an iteration in about 0.25 s. Simulations with group-specific blocking and Sybil enumeration show that G-Lox improves robustness over Lox- and rBridge-like baselines among systems that maintain broad issuance.
