Garbled Circuits (GC) is a foundational primitive for secure two-party computation (2PC). Garbled Private Information Retrieval (GPIR) is a GC technique for looking up a public array or database (DB) on a private index unknown to either player. GPIR immediately implies GC evaluation of functions implemented as a publicly known look-up table (LUT).
Garbled Circuits (GC) is a foundational primitive for secure two-party computation (2PC). Garbled Private Information Retrieval (GPIR) is a GC technique for looking up a public array or database (DB) on a private index unknown to either player. GPIR immediately implies GC evaluation of functions implemented as a publicly known look-up table (LUT).
Garbled Circuit (GC) is the main practical 2PC technique, yet despite great interest in its performance, GC notoriously resists improvement. Essentially, we only know how to evaluate GC functions gate-by-gate using encrypted truth tables; given input labels, the GC evaluator decrypts the corresponding output label. Interactive protocols enjoy more sophisticated techniques. For example, we can expose to a party a (masked) private value. The party can then perform useful local computation and feed the resulting cleartext value back into the MPC. Such techniques are not known to work for GC. We show that it is, in fact, possible to improve GC efficiency, while keeping its round complexity, by exposing masked private values to the evaluator. %without introducing rounds of communication. Our improvements use garbled one-hot encodings of values. By using this encoding we improve a number of interesting functions, e.g., matrix multiplication, integer multiplication, field element multiplication, field inverses and AES S-Boxes, integer exponents, and more. We systematize our approach by providing a framework for designing such GC modules. Our constructions are concretely efficient. E.g., we improve binary matrix multiplication inside GC by more than 6x in terms of communication and by more than 4x in terms of WAN wall-clock time. Our improvement circumvents an important GC lower bound and may open GC to further improvement.
Abstract. Garbled circuits is a cryptographic technique, which has been used among other things for the con-struction of two and three-party secure computation, private function evaluation and secure outsourcing. Garbling schemes is a primitive which formalizes the syntax and security properties of garbled circuits. We dene a gener-alization of garbling schemes called reactive garbling schemes. We consider functions and garbled functions taking multiple inputs and giving multiple outputs. Two garbled functions can be linked together: an encoded output of one garbled function can be transformed into an encoded input of the other garbled function without commu-nication between the parties. Reactive garbling schemes also allow partial evaluation of garbled functions even when only some of the encoded inputs are provided. It is possible to further evaluate the linked garbled functions when more garbled inputs become available. It is also possible to later garble more functions and link them to the ongoing garbled evaluation. We provide rigorous denitions for reactive garbling schemes. We dene a new notion of security for reactive garbling schemes called condentiality. We provide both simulation based and in-distinguishability based notions of security. We also show that the simulation based notion of security implies the indistinguishability based notion of security. We present an instantiation of reactive garbling schemes. We present an application of reactive garbling schemes to reactive two-party computation secure against a malicious adversary. We demonstrate how garbling schemes can be used to give abstract black-box descriptions and proof of several advanced applications of garbled circuits in the literature, including Minilego and Lindell’s forge-and-loose technique. 1
Garbled circuits is a cryptographic technique, which has been used among other things for the construction of two and threeparty secure computation, private function evaluation and secure outsourcing. Garbling schemes is a primitive which formalizes the syntax and security properties of garbled circuits. We define a generalization of garbling schemes called reactive garbling schemes. We consider functions and garbled functions taking multiple inputs and giving multiple outputs. Two garbled functions can be linked together: an encoded output of one garbled function can be transformed into an encoded input of the other garbled function without communication between the parties. Reactive garbling schemes also allow partial evaluation of garbled functions even when only some of the encoded inputs are provided. It is possible to further evaluate the linked garbled functions when more garbled inputs become available. It is also possible to later garble more functions and link them to the ongoing garbled evaluation. We provide rigorous definitions for reactive garbling schemes. We define a new notion of security for reactive garbling schemes called confidentiality. We provide both simulation based and indistinguishability based notions of security. We also show that the simulation based notion of security implies the indistinguishability based notion of security. We present an instantiation of reactive garbling schemes. We present an application of reactive garbling schemes to reactive two-party computation secure against a malicious, static adversary. We demonstrate how garbling schemes can be used to give abstract black-box descriptions and proof of several advanced applications of garbled circuits in the literature, including Minilego and Lindell’s forge-and-lose technique.
In this thesis we design and implement three approaches to provide efficient Secure Two- Party Computation (2PC) protocols that are secure in presence of malicious adversaries. The field of secure two-party computation enables two mutually distrusting parties, each with their own private input, to run a computation of an arbitrary program. At the end of the computation, nothing about their private input is leaked except for what can be learned from the output of the computation. The seminal work of Yao in 1986 provided the first proof that such a computation is feasible using the so called garbled circuits and given an Oblivious Transfer protocol. In this thesis, we follow Yao’s garbled circuit approach and focus on efficient protocols that are secure against malicious adversaries. We examine three different categories of optimization. In each category, we present a 2PC protocol, prove its security, and implement it as a proof of concept to demonstrate its running time in practice. Non-interactive Secure Computation: In this category, we offer the first implementation of a malicious 2PC protocol that reduces the required rounds of interaction between the two parties to only two rounds (i.e., each party sends only one message to the other party). Mixed Secure Two-Party Computation: In the second category we target big and complex programs and break them into subprograms. Then, we compute each subprogram using the most efficient 2PC protocol that is suitable for it. Finally, we securely connect these subprograms together. The goal is to reduce the overhead of the garbled circuits significantly, which results in a more efficient overall 2PC protocol for programs with specific properties. Secure Computation for Random Access Machine Programs: In the third category, we target programs that require random access to the memory (as opposed to a sequential access). We present the first efficient maliciously-secure protocol for computing such programs.
Abstract. We consider generic Garbled Circuit (GC)-based techniques for Secure Function Evaluation (SFE) in the semi-honest model. We describe efficient GC constructions for addition, subtraction, multiplication, and comparison functions. Our circuits for subtraction and comparison are approximately two times smaller (in terms of garbled tables) than previous constructions. This implies corresponding computation and communication improvements in SFE of functions using our efficient building blocks. The techniques rely on recently proposed “free XOR ” GC technique. Further, we present concrete and detailed improved GC protocols for the problem of secure integer comparison, and related problems of auctions, minimum selection, and minimal distance. Performance improvement comes both from building on our efficient basic blocks and several problemspecific GC optimizations. We provide precise cost evaluation of our constructions, which serves as a baseline for future protocols.
Applying cut-and-choose techniques to Yao’s garbled circuit protocol has been a promising approach for designing efficient Two-Party Computation (2PC) with malicious and covert security, as is evident from various optimizations and software implementations in the recent years. We revisit the security and efficiency properties of this popular approach and propose alternative constructions and definitions that are more suitable for use in practice. • We design an efficient fully-secure malicious 2PC protocol for two-output functions that only requires O(t|C|) symmetric-key operations (with small constant factors) where |C | is the circuit size and t is a statistical security parameter. This is essentially the optimal complexity for protocols based on cut-and-choose, resolving a main question left open by the previous work on the subject. Our protocol utilizes novel techniques for enforcing garbler’s input consistency and handling twooutput functions that are more efficient than all prior solutions. • Motivated by the goal of eliminating the all-or-nothing nature of 2PC with covert security (that privacy and correctness are fully compromised if the adversary is not caught in the challenge phase), we propose a new security definition for 2PC that strengthens the guarantees provided by the standard