The Future of Fully Homomorphic Encryption System: from a Storage I/O Perspective

TL;DR

This paper examines how storage I/O bottlenecks significantly impact the performance of Fully Homomorphic Encryption systems, highlighting the need for optimized I/O solutions to enhance practical deployment.

Contribution

It provides an analysis of storage I/O effects on FHE performance and summarizes key lessons, addressing an understudied aspect of FHE deployment.

Findings

Storage I/O can degrade ASIC performance by up to 357 times.

GPU performance can be reduced by up to 22 times due to I/O bottlenecks.

Highlights the importance of optimizing storage I/O for practical FHE applications.

Abstract

Fully Homomorphic Encryption (FHE) allows computations to be performed on encrypted data, significantly enhancing user privacy. However, the I/O challenges associated with deploying FHE applications remains understudied. We analyze the impact of storage I/O on the performance of FHE applications and summarize key lessons from the status quo. Key results include that storage I/O can degrade the performance of ASICs by as much as 357$\times$ and reduce GPUs performance by up to 22$\times$.