Towards Interference-Resilient Multi-Tenant Microservices via Spatio-Temporal Models of Self-Configuration

Microservices offer fine-grained scalability and facilitate easier deployment on shared resources. On the other hand, simultaneous utilization of these resources may result in obscure interdependencies and complicated interference patterns that spread abnormalities in performance throughout different applications. To address this challenge, we have developed a new graph-based formalism, the Spatio-Temporal Interference Graph (STIG), which maps these interference patterns and simulates their dynamics. My PhD thesis aims to shed insights on performance anomalies caused by common irregularities (i.e. interference) in resource consumption and develop proactive strategies to reduce interference in the future, achieve high accuracy, and produce a low number of false alarms through self-adaptive mechanisms.