괴델상 2022 수상자 초청강연

 

Abstract: We present SimplePIR, the fastest private information retrieval (PIR) scheme known to date. SimplePIR is a single-server PIR scheme, whose security holds under the learning-with-errors assumption. To answer a client’s PIR query, the SimplePIR server performs one 32-bit multiplication and one 32-bit addition per database byte. SimplePIR achieves 6.5 GB/s/core server throughput, which is 7% faster than the fastest two-server PIR schemes (which require non-colluding servers). SimplePIR has relatively large communication costs: to make queries to a 1 GB database, the client must download a 124 MB “hint” about the database contents; thereafter, the client may make an unbounded number of queries, each requiring 242 KB of communication. We present a second single-server scheme, DoublePIR, that shrinks the hint to 16 MB at the cost of slightly higher per-query communication (345 KB) and slightly lower throughput (5.2 GB/s/core). Finally, we apply our PIR schemes, together with a new data structure for approximate set membership, to the problem of private auditing in Certificate Transparency. We achieve a strictly stronger notion of privacy than Google Chrome’s current approach with a modest, 13x larger communication overhead.