Building and Analyzing Arbitrary Cloud-based Networks

Abstract

Cloud networking has become a critical engineering discipline as organisations increasingly rely on public cloud platforms to host distributed applications and interconnect services at scale. Despite widespread adoption, empirical evaluation of cloud network architectures at the topology level remains limited. Most existing research benchmarks cloud platforms at the virtual machine or application layer, leaving a gap in experimentally validated guidance for engineers designing enterprise-grade network topologies. This thesis addresses that gap by designing, implementing, and experimentally evaluating hub-and-spoke network architectures on Amazon Web Services and Microsoft Azure using real, live cloud infrastructure. All environments were provisioned through Terraform Infrastructure as Code, with the complete codebase and test scripts publicly available at https://github.com/BlossomA/cloudnetworks.