Arquitetura de nuvem privada para ambientes acadêmicos heterogêneos: um estudo de planejamento e reaproveitamento de hardware com Proxmox VE

Abstract

This study addresses the operational challenges arising from technological heterogeneity and hardware obsolescence in the shared computer laboratories at UFC Campus de Russas. The individualized installation and maintenance of software in these environments lead to workload overhead and a lack of standardization, which negatively impacts the multidisciplinary academic experience. The objective of this study is to design the logical architecture and evaluate the technical viability of a private cloud based on the Proxmox VE platform, focusing on the reuse of legacy hardware comprising Intel Core i5-3470 processors, 8 GiB of RAM, and 120 GB SSD units. The methodology is characterized by a hypothetic-deductive approach, utilizing a case study strategy and a quantitative analytical modeling based on technical specifications for resource sizing. The analytical results indicate that virtualization through LXC containers stably sustains a simultaneous load of 30 users under lightweight development profiles. However, the memory density modeling points out severe limitations in the useful physical RAM ceiling per node (6.5 GiB), indicating the unfeasibility of full virtualization scenarios (KVM) with heavy graphical interfaces executed under synchronous maximum capacity. It is concluded that the proposed architectural model supports the technical viability of the private cloud for the campus, converting physical constraints into a centralized and resilient infrastructure, while delegating the empirical validation of the cluster performance through physical stress testing as a recommendation for future work.