If you have any problems related to the accessibility of any content (or if you want to request that a specific publication be accessible), please contact us at firstname.lastname@example.org.
Private Multi-Cloud Architectural Solutions for NRDC Data Streaming Services
AltmetricsView Usage Statistics
Cloud computing has experienced huge growth in both commercial and research applications in recent years. In order to maximize computing efficiency, virtual machines are collocated on the same processors. In this research, OpenStack was studied as a potential platform to migrate the aging Nevada Research Data Center infrastructure that was built on Microsoft Server 2012. This data center hosted virtual machines that ingested data from remote towers and made the data available to environmental researchers through a web portal. The NRDC ultimately had a storage failure and a temporary hybrid infrastructure was set up to prevent data loss. Comparisons of the visual management interfaces, licensing cost, and license type were made between OpenStack and Windows Server. The Windows Server virtualization interface was susceptible to information overload and lacked the security features included with the OpenStack interface. Performance comparisons were done on instance launching and deletion for the OpenStack test cluster to study the effects image and instance resources. There were differences found for real and sys timings for launch timings using the OpenStack Command line interface (CLI). However, no differences were found between the two instance types when launched or deleted using the Python application programming interface (API). These timings could have possibly been skewed by several outliers. Performance differences between the CLI and Python API was also tested. The Python API performed better than the CLI for real, sys, and user API call timings, but worse for the backend spawn, build, delete on hypervisor, and deallocation of network resources timings.