Optimization of the Cloud-Native Infrastructure using Artificial Intelligence

Sammanfattning: To test Cloud RAN applications, such as the virtual distributed unit (vDU) and centralized virtual unit (vCU), a test environment is required, commonly known as a “test bed” or “test channel”. This test bed comprises various cloudnative infrastructures, including different hardware and software components. Each test bed possesses distinct capacities for testing various features, leading to varying costs. With the increasing number of cloud applications, additional test beds are necessary to ensure thorough testing before releasing these applications to the market. To optimize the creation process of a Cloud-native test bed, leveraging artificial intelligence and machine learning approaches can be beneficial. This thesis presents, applies, and evaluates an AI-based approach for optimizing the construction of Cloud-native test beds. The proposed solution’s feasibility is assessed through an empirical evaluation conducted in the Telecom domain at Ericsson AB in Sweden.