PPRL(Privacy-Preserving Record Linkage)

PPRL solves the problem of matching records across organizations that have legal or regulatory reasons not to exchange raw data. A bank and an insurer want to know which customers overlap; neither can hand the other their full customer list.

The dominant PPRL technique encodes each record into a Cryptographic Linkage Key (CLK) — a Bloom filter generated from the record's identifying fields using a shared schema and a shared salt. Both sides exchange CLKs (not raw data) with a linkage broker, which computes Bloom-filter intersections and returns which CLK IDs match. Each side resolves their own CLK IDs back to local records.

Critically, the salt is shared between the two organizations and *not* known to the broker — otherwise the broker could attack the Bloom filter and recover raw values. PPRL is the gold standard for cross-tenant matching in regulated industries (health, finance, government).