Broker-based private matching

Private matching solutions allow two parties to find common data elements over their own datasets without revealing any additional private information. We propose a new concept involving an intermediate entity in the private matching process: we consider the problem of broker-based private matching where end-entities do not interact with each other but communicate through a third entity, namely the Broker, which only discovers the number of matching elements. Although introducing this third entity enables a complete decoupling between end-entities (which may even not know each other), this advantage comes at the cost of higher exposure in terms of privacy and security. After defining the security requirements dedicated to this new concept, we propose a complete solution which combines searchable encryption techniques together with counting Bloom filters to preserve the privacy of end-entities and provide the proof of the matching correctness, respectively.