Cloud-based secure health monitoring: Optimizing fully-homomorphic encryption for streaming algorithms

There are many incentives for healthcare providers to shift their datacenters to the cloud. However, privacy of patient health information is a major concern when processing medical data off-site. One possible solution is the use of Fully Homomorphic Encryption (FHE), but this solution is too slow for most applications. We present a technique that increases efficiency and parallelism for certain algorithms under FHE. Through simulations, we demonstrate that our method yields about 20x speedup in a sample application. This is a significant step towards practical FHE-based medical remote monitoring.