Towards Reconstructing Multi-Step Cyber Attacks in Modern Cloud Environments with Tripwires

Abstract

Rapidly-changing cloud environments that consist of heavily interconnected components are difficult to secure. Existing solutions often try to correlate many weak indicators to identify and reconstruct multi-step cyber attacks. The lack of a true, causal link between most of these indicators still leaves administrators with a lot of false-positives to browse through. We argue that cyber deception can improve the precision of attack detection systems, if used in a structured, and automatic way, i.e., in the form of socalled tripwires that ultimately span an attack graph, which assists attack reconstruction algorithms. This paper proposes an idea for a framework that combines cyber deception, automatic tripwire injection and attack graphs, which eventually enables us to reconstruct multi-step cyber attacks in modern cloud environments. CCS CONCEPTS• Security and privacy → Intrusion/anomaly detection and malware mitigation; Web application security; Network security.

An optimal defensive deception framework for the container‐based cloud with deep reinforcement learning

“…Given the frequent updating and scaling of microservice, active applications and their mappings to computing nodes changes constantly [11]. Hence, it is essential for the placement strategy of decoy to adapt to the highly dynamic cloud system [30].…”