Abstract Cloud computing has evolved into a globally distributed, partially autonomous substrate that effectively orchestrates much of the “invisible” Internet. While extensive work has addressed performance, reliability, and cost optimization, significantly less attention has been paid to the explicit modeling of autonomy, architectural emergence, and system-level value creation in multi-tenant cloud environments. This paper introduces EVO-ORCH (Evolutionary Value-Oriented Orchestrator), a hierarchical orchestration framework that treats cloud services as autonomous agents operating under global architectural and economic constraints. A formal notion of Emergent Value Index (EVI) is proposed, combining availability, latency, resource efficiency, and adaptation behavior into a single quantitative metric. EVO-ORCH leverages adaptive profiling and learning-based control to select orchestration actions (placement, scaling, routing) that maximize EVI while satisfying service-level objectives. A hypothetical yet defensible experimental study on a simulated multi-tenant cloud shows that EVO-ORCH can improve EVI by up to 18.6% compared with a baseline Kubernetes-style orchestrator, reduce SLO violation rate by 27.3%, and maintain control-plane overhead below 6.4% of total resources. Complexity analysis and architectural discussion indicate that the proposed framework is compatible with existing cloud control planes. The results suggest that value-oriented, autonomy-aware orchestration can provide a meaningful step towards more self-managing cloud infrastructures in practice.
Jangale et al. (Sat,) studied this question.