Distributed and Intelligent API Mediation Service for Enterprise-Grade Hybrid-Multicloud Computing

In an enterprise-grade hybrid-multicloud computing environment, capability-providing as-a-service endpoints (or aaS-endpoints) can be deployed across diverse computing platforms, e.g., public clouds and on-prem enterprise private clouds. To ensure a seamless, unified, and enterprise-compliant acquisition of the capabilities by client applications, the presence of a cross-cloud API mediation service is crucial. However, as the number and heterogeneity of aaS-endpoints increase, delivering the API mediation service at scale becomes increasingly costly. This paper presents a robust approach to API service mediation in enterprise-grade hybrid-multicloud computing environments. It tackles the challenges, offering a distributed architecture comprising dynamically composed managed microservices, microservice zones, intelligent endpoint selection, and adaptive statistical learning (aiming to exploit localities in performance history of aaS-endpoint invocations and to facilitate adding or removing active aaS-endpoints). The successful reference implementation and $24\mathrm{x}7\mathrm{x}365$ delivery in real-world settings of the approach validate its efficacy as a practical solution for API service mediation.