作者: Domenico Cotroneo , Luigi De Simone , Roberto Natella
DOI: 10.1109/TNSM.2017.2733042
关键词:
摘要: Network function virtualization (NFV) envisions the use of cloud computing and technology to reduce costs innovate network services. However, this paradigm shift poses question whether NFV will be able fulfill strict performance dependability objectives required by regulations customers. Thus, we propose a benchmark support providers at making informed decisions about which virtualization, management, application-level solutions can achieve best dependability. We define in detail cases, measures, faults injected. Moreover, present benchmarking case study on two alternative, production-grade solutions, namely VMware ESXi/vSphere (hypervisor-based) Linux/Docker (container-based), deploy an NFV-oriented IMS system. Despite promise higher manageability, our experiments suggest that container-based configuration less dependable than hypervisor-based one, point out designers should address improve