Privacy-Preserving Data Loss Prevention via Encrypted Set Intersection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data loss prevention systems struggle to accurately detect sensitive data leakage while preserving privacy, often resulting in missed detections and noise due to the inability to determine sensitive data contextually and the risk of disclosing sensitive information to third-party hosts.

Innovation Solution

A computer-implemented method and system that employs a privacy-preserving set intersection protocol, where data communications are encrypted and analyzed externally to a firewall, allowing for matching between encrypted databases and communications without revealing cleartext data, thus enabling accurate detection and action on data leakage while maintaining privacy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If an enterprise uses external hosts to perform data loss prevention analysis, then detection capability is improved, but sensitive data privacy is compromised due to disclosure to third parties

Engineering Contradiction:
Improvedetection capabilityVSAvoidsensitive data privacy
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent introduces an intermediary encryption layer between the enterprise and external DLP hosts. Data is encrypted before leaving the enterprise firewall, and the external host performs analysis on encrypted data. This intermediary encryption mechanism allows external hosts to perform detection without directly accessing plaintext sensitive data, thus maintaining privacy while improving detection capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transforms the data from its original plaintext state to an encrypted state before transmission to external hosts. This parameter change (from plaintext to ciphertext) allows the data to be analyzed for detection purposes while preventing direct access to the actual sensitive information, thereby resolving the contradiction between detection capability and privacy protection.

Inventive Principle:
Principle #35Parameter changes

2Speed

If pattern matching algorithms are used to detect sensitive data, then detection speed is improved, but accuracy deteriorates due to high noise levels and missed detections

Engineering Contradiction:
Improvedetection speedVSAvoiddetection accuracy
Core Design Contradiction:
SpeedVSMeasurement precision

Solution Approach 1:

The patent performs preliminary encryption and context analysis before data leaves the enterprise. By preparing and encrypting data in advance with contextual information preserved, the system enables faster external processing while maintaining accuracy through pre-established encryption patterns that retain detectable features for matching algorithms.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If enterprises provide sensitive data to third-party DLP hosts, then detection capability is improved, but security risk increases due to loss of control over data

Engineering Contradiction:
Improvedetection capabilityVSAvoidsecurity risk
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

Encryption serves as an intermediary protective layer that allows data to be shared with third-party hosts for detection purposes while maintaining security control. The encryption mechanism ensures that even though data is transmitted externally, the enterprise retains control through key management, preventing unauthorized access or misuse by third parties.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3338401B1Systems and methods for data loss prevention while preserving privacy
Publication Date: 2020.03.25 GOOGLE LLC
  • EP3338401B1 patent drawingFigure 1
  • EP3338401B1 patent drawingFigure 2
  • EP3338401B1 patent drawingFigure 3

AI summary

Systems and methods for data loss prevention while preserving privacy include receiving a data communication originating from an online account of a user device associated with the enterprise; performing an analysis to determine that the data communication is a candidate data leakage, based on a context of the data communication; encrypting the data communication, and providing the host-encrypted data communication to a software agent at the enterprise; receiving a software agent-encrypted database of enterprise communication and the host-encrypted data communication, re-encrypted by the software agent; decrypting a host-encrypted aspect of the re-encrypted data communication to generate a software agent-encrypted data communication; performing a matching operation to determine whether a match exists between the encrypted database and the software agent-encrypted data communication; if the match exists, reporting the match to the software agent; else, the host taking an action.