MPC with Self-Adjusting Security Levels Based on Real-Time Threat Intelligence

Most existing MPC frameworks, such as those discussed by Deevi (2023) or Anagreh & Laud (2024), statically set their security models and parameters at deployment. However, the threat landscape is not static: attacks may spike, insider threats may be detected, or external intelligence may indicate new vulnerabilities. This research proposes MPC protocols that monitor real-time threat feeds (e.g., from SIEMs, security analytics, or collaborative sharing), and then autonomously “harden” or “relax” their internal operations as needed—e.g., by increasing the number of computation rounds, switching to stronger cryptographic primitives, or enabling additional verification steps. This self-adjusting paradigm would be a substantial step beyond today’s “set-and-forget” security, offering resilience and efficiency by only invoking costly protections when justified by context.

References:

1. Federated Machine Learning Across Hybrid Clouds: Balancing Security and Privacy. Sri Ramya Deevi (2023). International Journal of Computing and Engineering.
2. Security Proof of Single-Source Shortest Distance Protocols Built on Secure Multiparty Computation Protocols. Mohammad Anagreh, Peeter Laud (2024). Cryptogr..