Verizon has completed one of the industry’s largest known Network Function Virtualization OpenStack cloud deployments across five of its U.S. data centers. The NFV project, which began in 2015, created a production design based on a core and pod architecture that provides the hyperscale capabilities and flexibility necessary to meet the company’s complex network requirements.
Verizon worked with Big Switch Networks, Dell and Red Hat to develop the OpenStack pod-based design that went from concept to deployment of more than 50 racks in five production data centers in less than nine months.
Deployments are currently in progress in additional domestic data center and aggregation sites, with international locations to be deployed over the next several months. The design also will be adopted in edge network sites by the end of the year.
“Building on our history of innovation, this NFV project is another step in building Verizon’s next-generation network – with implications for the industry,” said Adam Koeppe, vice president, network technology planning, Verizon. “New and emerging applications are highlighting the need for collaborative research and development in technologies like NFV. We consider this achievement to be foundational for building the Verizon cloud that serves our customers’ needs anywhere, anytime, any app.”
Executives from Verizon, Big Switch, Dell and Red Hat will speak about the project at the OpenStack Summit in Austin, Texas, April 25 – 29.
Red Hat Openstack
To validate the resiliency of the NFV pod design at scale, the collaborators constructed and tested large scale test beds mirroring the production design, leveraging the open source community to ultimately deliver a high quality, validated NFV pod architecture.
The project is based on OpenStack with Red Hat Ceph Storage and a spine-leaf fabric for each pod controlled through a Neutron plugin to Red Hat OpenStack Platform. The multi-vendor deployment leverages Big Switch’s Big Cloud Fabric for SDN controller software managing Dell switches, which are orchestrated by Red Hat OpenStack Platform.
To meet the challenges of a large scale NFV deployment, the companies worked together to design a solution that addressed five key needs identified early in the project:
- Resiliency at Scale – The design followed a hyperscale-inspired “core and pod” approach with a 12-rack pod design replicated at data centers across the United States.
- No Bandwidth Bottlenecks – A modern leaf-spine Clos design, using centralized SDN control designed to take the network from spine to leaf to vSwitch and avoid bandwidth bottlenecks.
- Logical Network Design Flexibility – The pod design would accommodate unique NFV workloads with unique logical network requirements that share the same physical leaf/spine fabric and vSwitches.
- Reduced Operational Complexity – Operational complexity would be reduced through a simplified lifecycle management of the network control systems relative to the OpenStack control systems.
- Integrated Security and Visibility – The NFV Pod is designed to be compliant and secured against intrusions and other threats, monitoring fabric was used to monitor intra-pod traffic and inline traffic.