Adaptive Observability for Forensic-Ready Microservice Systems

Monteiro, Davi; Yu, Yijun; Zisman, Andrea and Nuseibeh, Bashar (2023). Adaptive Observability for Forensic-Ready Microservice Systems. IEEE Transactions on Services Computing, 16(5) pp. 3196–3209.

DOI: https://doi.org/10.1109/tsc.2023.3290474

Abstract

Microservice-based applications may include multiple instances of microservices running on containerised infrastructures. These infrastructures pose challenges to digital investigations of security incidents because digital evidence can be destroyed when containers are terminated. Observability techniques are used to facilitate the investigation of incidents in microservice systems. However, existing observability approaches do not address security incidents when there is a need to perform digital forensic investigations. Furthermore, approaches to proactively support digital forensic investigations are limited to security incidents that are known a priori. In this paper, we propose an adaptive observability approach based on game theory. The approach addresses the challenge of implementing forensic-ready microservice systems while considering uncertainties in security incidents. Our approach provides evidence collection capabilities for microservice systems and continually adapts to improve the forensic readiness of microservices. Specifically, the approach uses game theory to model and reason about the interactions between users and microservices, determining the optimal time and manner for observing microservices before the occurrence of security incidents. The performance of the approach has been assessed and compared with other observability approaches. Results of the evaluation indicate that adaptive observability outperforms other observability approaches, with improvements ranging from 3.1% up to 42.50%.

Viewing alternatives

Metrics

Item Actions

Export

You can export this page using these formats Digital Object Identifier - DOI