Architects designing NFV deployments focus most of their time on providing robust network, compute and automation layers, while considering the challenges of operating at scale.
What is often neglected during the planning process are the real world requirements for persistent storage to provide their services. As architects move from designing customized to designing standardized and automated stacks, it will become more important to provide a flexible, yet easy to maintain, storage architecture that can guarantee performance for each application in a NFV deployment.
Looking forward, network functions are going to hosted in different formats: physical, virtualized, containerized, potentially even from the public cloud. How can we make sure all these flavors can coexist all together, without forcing us to re-architect our current storage setups?
