Data centers are a critical part of the Internet infrastructure. In fact, most of the relevant online services are hosted in a data center. Data center networks are complex, since they are characterized by a high density architecture and by a high level of redundancy. Fat tree topologies are currently the most used in hyperscale data centers. Performing tests in such topologies would be unfeasible, because of the high costs of the required equipment and due to the involvement of human resources. This would limit the automation and reproducibility of tests, leading to a more error-prone testing pipeline. This paper presents VFTGen, a tool that, leveraging on the virtualization and the Software Defined Data Center concepts, automatically builds, deploys and configures arbitrary fat tree topologies in a virtual environment. We demonstrate the ease of use of the tool and its value as a support to the study or the development of networking protocols for fat trees.
Caiazzi, T., Scazzariello, M., Ariemma, L. (2021). VFTGen: A Tool to Perform Experiments in Virtual Fat Tree Topologies. In Proceedings of the IM 2021 - 2021 IFIP/IEEE International Symposium on Integrated Network Management (pp.718-719). Institute of Electrical and Electronics Engineers Inc..
VFTGen: A Tool to Perform Experiments in Virtual Fat Tree Topologies
Caiazzi T.;Scazzariello M.;Ariemma L.
2021-01-01
Abstract
Data centers are a critical part of the Internet infrastructure. In fact, most of the relevant online services are hosted in a data center. Data center networks are complex, since they are characterized by a high density architecture and by a high level of redundancy. Fat tree topologies are currently the most used in hyperscale data centers. Performing tests in such topologies would be unfeasible, because of the high costs of the required equipment and due to the involvement of human resources. This would limit the automation and reproducibility of tests, leading to a more error-prone testing pipeline. This paper presents VFTGen, a tool that, leveraging on the virtualization and the Software Defined Data Center concepts, automatically builds, deploys and configures arbitrary fat tree topologies in a virtual environment. We demonstrate the ease of use of the tool and its value as a support to the study or the development of networking protocols for fat trees.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.