Plinius: Secure and Persistent Machine Learning Model Training

TL;DR

PLINIUS is a secure, high-performance machine learning framework that leverages persistent memory and Intel SGX enclaves to ensure privacy, integrity, and fault tolerance during model training.

Contribution

It introduces a novel mirroring mechanism using persistent memory for encrypted, near-instantaneous recovery and secure training of ML models within SGX enclaves.

Findings

PLINIUS is 3.2x faster for saving models compared to disk-based systems.

PLINIUS achieves 3.7x faster model restoration on real hardware.

The framework provides robust security and fault tolerance for ML training.

Abstract

With the increasing popularity of cloud based machine learning (ML) techniques there comes a need for privacy and integrity guarantees for ML data. In addition, the significant scalability challenges faced by DRAM coupled with the high access-times of secondary storage represent a huge performance bottleneck for ML systems. While solutions exist to tackle the security aspect, performance remains an issue. Persistent memory (PM) is resilient to power loss (unlike DRAM), provides fast and fine-granular access to memory (unlike disk storage) and has latency and bandwidth close to DRAM (in the order of ns and GB/s, respectively). We present PLINIUS, a ML framework using Intel SGX enclaves for secure training of ML models and PM for fault tolerance guarantees. PLINIUS uses a novel mirroring mechanism to create and maintain (i) encrypted mirror copies of ML models on PM, and (ii) encrypted training data in byte-addressable PM, for near-instantaneous data recovery after a system failure. Compared to disk-based checkpointing systems, PLINIUS is 3.2x and 3.7x faster respectively for saving and restoring models on real PM hardware, achieving robust and secure ML model training in SGX enclaves.