Garbled AND Gate Pool for Secure Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current secure computation services face limitations in scalability and efficiency, particularly in handling reactive computations like RAM-based computations against active adversaries, and often require substantial offline processing, making them unsuitable for large computations.
Innovation Solution
A computer-implemented method using a pool of garbled AND gates, where garbled circuits and cut-and-choose techniques enable secure computation by randomly selecting and evaluating garbled gates, allowing for efficient execution of logical AND gates without incurring long delays or large storage demands, and employing LEGO protocols for scalability and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional secure computation approaches are used, then security against active adversaries is achieved, but scalability and efficiency deteriorate due to substantial offline processing requirements
Solution Approach 1:
The computation is divided into multiple independent garbled circuits that can be processed in parallel. Each circuit handles a portion of the overall computation, allowing the system to scale by adding more circuits without increasing the offline processing burden for each individual circuit.
Solution Approach 2:
Garbled circuits are pre-generated and stored in a pool during an offline phase. During the online evaluation phase, circuits are randomly selected from this pre-generated pool, eliminating the need for substantial offline processing during actual computation execution.
2Loss of time
If offline processing is reduced for on-demand service, then responsiveness improves, but computation security and correctness may deteriorate
Solution Approach 1:
Garbled circuits are generated in advance during an offline setup phase and stored in a pool. This preliminary action allows the online phase to quickly select and evaluate pre-validated circuits without performing extensive offline processing during on-demand execution.
Solution Approach 2:
The system incorporates verification mechanisms where the evaluating party can check the correctness of garbled circuits during evaluation. This feedback loop ensures computation correctness is maintained even with reduced offline processing.
3Productivity
If more garbled circuits are pre-generated for scalability, then service capacity increases, but storage requirements and resource usage increase
Solution Approach 1:
A single pool of garbled circuits serves multiple computation tasks and different evaluating parties. The same pre-generated circuits can be reused across various on-demand secure computation requests, reducing the need to generate and store separate circuits for each specific task.
Solution Approach 2:
Garbled circuits are randomly selected from the pool for evaluation and then discarded after use. This allows the system to maintain a manageable pool size while providing scalable service, as circuits are reused rather than continuously accumulated.
Data Source
AI summary
Technologies are disclosed herein for running a long-term on-demand service for executing actively-secure computations. A function circuit may be represented as a stream of buckets, in which each bucket represents a logical AND gate. A pool having a plurality of garbled AND gates is generated. Garbled AND gates are randomly selected from the pool for placement in one of the buckets. An output for the bucket is determined by an evaluation of the selected garbled AND gates. The output represents an execution of the logical AND gate. The determined output is applied as a parameter in a secure protocol.