Advancements in Runtime Security and AI/ML Optimization

The field of cloud and cloud-native computing is seeing significant advancements in runtime security and AI/ML optimization. Researchers are exploring new ways to enhance security in containerized and virtualized environments, such as using eBPF technology to monitor and enforce policies. Additionally, there is a growing focus on optimizing AI/ML workflows, including the development of novel systems for declarative factorization of AI/ML inferences over joins. Noteworthy papers include: eBPF-PATROL, which introduces an extensible lightweight runtime security agent for detecting and preventing real-time boundary violations. InferF, which proposes a declarative system for factorizing AI/ML inferences over multi-way joins, achieving up to 11.3x speedups. Securing the Model Context Protocol, which highlights new security risks and proposes practical controls for securing dynamic, user-driven agent systems.

Sources

eBPF-PATROL: Protective Agent for Threat Recognition and Overreach Limitation using eBPF in Containerized and Virtualized Environments

InferF: Declarative Factorization of AI/ML Inferences over Joins

Securing the Model Context Protocol (MCP): Risks, Controls, and Governance

AI/ML Model Cards in Edge AI Cyberinfrastructure: towards Agentic AI

Built with on top of