Federated CPU Tracker
Modern organizations often need to monitor CPU usage across their network to optimize performance and resource allocation. However, this presents a significant security challenge: raw CPU data can be exploited by sophisticated attacks like Hertzbleed and PLATYPUS to extract sensitive information, including cryptographic keys.
To address this challenge, we've developed a solution that uses Differential privacy - a mathematical framework that protects sensitive data by adding carefully calibrated noise. This approach strikes an optimal balance: it prevents malicious actors from accessing exact CPU measurements while preserving the statistical utility of the aggregated data for legitimate monitoring purposes.
Using SyftBox's API framework, we'll build a system that:
- Collects CPU usage data securely
- Applies differential privacy protection automatically
- Enables safe sharing of protected metrics across the SyftBox network
📄️ Part I: The Member API
A CPU monitoring client with SyftBox and Differential Privacy
📄️ Part II: The Aggregator API
A privacy-preserving CPU monitoring aggregator with SyftBox.