UMBC Cyber Defense Lab
SPARCLE: Practical Homomorphic Encryption
Russ Fink
Senior Scientist
Johns Hopkins University / Applied Physics Laboratory
12:00–1:00pm Friday, April 27, 2018, ITE 237, UMBC
In the newly coined Privacy Age, researchers are building systems with homomorphic algorithms that enable “never decrypt” operations on sensitive data in applications such as computational private information retrieval (cPIR). The trouble is, the leading algorithms incur significant computational and space challenges, relegating them mainly to large cloud computing platforms. We have invented a special-purpose, ring-homomorphic (aka, “fully homomorphic”) algorithm that, owing to some specializing assumptions, trades general-purpose cryptographic utility for linear performance in speed and space.
We will present the cryptosystem and discuss several current challenges. We will also throw in a fun, simple, tactile concept demonstration of PIR for those just generally curious about what all this is, hopefully demystifying how you can enable a server to search for something without knowing what it’s looking for, and without knowing what (if any) results it found.
Russ Fink (UMBC ’10) is a senior scientist at the Johns Hopkins University / Applied Physics Laboratory. His current research interests include private information retrieval, applied cryptography, and cyber security.