Using Digital Twins for Software Change Risk Assessment Toward Proactive AIOps.

The increasing structural and behavioural complexity of modern IT systems and environments (IT-Sys|Envs) calls for adopting automated fault anticipation and forecasting mechanisms to mitigate risks, limit system disturbances and damage, and improve problem resolution readiness. Preventing unexpected or undesired ITSys| Envs behaviour is crucial not only to maximize customer value and satisfaction but also to fulfill strict service-level agreements. Given the constant need for operational changes in today’s dynamic IT-Sys|Envs as a response to evolving user requirements and expectations, DevOps teams face several challenges and uncertainties to limit the potential introduction of faults, foresee their occurrence, and comprehend their associated risks. While the accomplishments attained in the application of artificial intelligence to the operation of IT-Sys|Envs (i.e., AIOps) show promise, further research is necessary to facilitate the transition from reactive AIOps to its anticipated proactive, and risk-focused, manifestation. This involves advancing AIOps conceptualization and tooling for enabling improved software fault prognosis and remediation, augmented risk management, and enhanced explainability. We describe our vision for proactive AIOps and report our progress toward its viable realization through the application of relevant notions from the revolutionary concept of Digital Twins. Moreover, we discuss challenges and opportunities regarding this promising integration and its impact on the future of software development and operation life cycles.