The OPNFV Summit will feature a Proof of Concept (PoC) Zone in the Technology Showcase highlighting industry projects and network operator use case demos. The feature real life applications of open source technologies done in partnership with service providers.
Confirmed PoCs include:
OPNFV Community Demo: Virtual Central Office (Cumulus, Ericsson, F5, Inocybe, Lenovo, Mellanox, Nokia, Red Hat)
Central Office Virtualization is a hot topic for Telecommunication Providers. While CORD (Central Office Re-Architected as a Data center) has been demonstrated before, However, Telcos continue to be interested in building the virtualization architecture of Central Offices using OpenDaylight SDN controller. This presentation and POC explains the need for this reference architecture, explains the architecture details as discussed and agreed upon by several Telcos and vendors using open source hardware and software components including OpenDaylight SDN controller, Openstack, open source orchestration software.The POC shows Residential services model on the VCO architecture with onboarding of VNFs, subscribers and delivering content and services.
Achieving Global Scale VNF High Availability with Multi-Site (Huawei Technologies, Orange, China Mobile)
In this PoC, we will demonstrate achieving global scale VNF High Availability with OPNFV Multi-Site based on Openstack Tricircle. We will use cloud based video conferencing and IMS as killer VNF apps to introduce the architecture and technical solutions, demonstrate design patterns and capabilities, and share HA service deployment considerations from Tier-1 operator perspective. Additional use case examples may be added to this PoC based on community feedback and participation.
Automatic integration, testing and certification system (China Mobile, Huawei)
As Operators expand their deployment of NFV, automatic integration, testing and compliance certification become more and more important. In this PoC, we would like to share our experience and progress in the China Mobile OPNFV Testlab on an automatic system of integration, testing and certification. This system takes fully use of OPNFV opensource tools, including installers such as Compass, testing such as Functest and Yardstick, compliance testing such as Dovetail. Such automatic system extremely decreases the human cost of Operators when deploying and testing the NFV cloud before large scale deployment.
Building the Carrier Grade NFV Infrastructure(China Telecom, Inspur)
- China Telecom NFVI Test Bed: The two sets of HA solutions UIVIM1 + RAID UI and UIVIM2 + CephU integrating VNF, MANO, OSS, EMS and other major components have been run in multiple sets of NFVI test beds
- Multi-Vendor VIM & Cloud Validation: Evaluation test of multiple VIM commercial products have been carried out on test beds and VNF network element deployment and docking work smoothly while MANO compatibility verification is still in the progress
High performance data plane networking in containers (China Mobile, Intel, ZTE Mobile)
A major advancement in cloud native technology is motivating the giant telecom industries to transform from VM based infrastructure to Container Orchestration Engines like Kubernetes. This has given rise to the need to support multi interface pods. Isolation of network traffic into different network planes such as management, control and media is a requisite feature in the telecom industry. Kubernetes currently has provision for a single network interface exposed inside the containers of a pod. Traditional container networking is based on kernel networking, which cannot provide a satisfactory performance for VNFs. This PoC proposes user space networking to address this problem. Through the use of the Multus and DPDK CNI plugins, we are achieving the high performance data plane networking with ultra-low latency in Kubernetes.
NFV Infrastructure Maintenance Automation by OPNFV Doctor (NEC Corporation, NTT DOCOMO, INC., NTT Corporation, Nokia, ZTE)
As real time evacuation is not possible for most telecom nodes, these are deployed in active-standby pairs. It was not possible so far to empty the active node automatically without any service downtime, when the active node itself becomes a maintenance candidate. This demo will show that OPNFV Doctor standard extensions in OpenStack can now automatically empty an active node without any service downtime by triggering an active to standby switchover when maintenance instructions are received. This reduces many operational tasks that an operator has so far manually performed during maintenance.
Run OPNFV, OPEN-O and Containerized vEPC together on Scorpio Multi-Node Server (China Mobile, Huawei, Intel, Mirantis, ODCC, Red Hat, ZTE)
- Introduction of Scorpio Multi-node server
- Deploy OPNFV Danube on Scorpio Multi-node server
- Deploy OPEN-O on Scorpio Multi-node server
- Deploy OpenRetriever on Scorpio Multi-node server
- Health check of whole system
- Workload of whole system
Scenario-oriented NFVO Demo (China Telecom)
This POC will show the procedure of utilising NFVO in IP network.
- Network cloud resource management
- vBRAS instance instantiation and configuration
- vFW lifecycle managment
- Network service orchestration
- VNF instance monitoring, warning and logging
VNF Onboarding and Service Creation in CORD (AT&T, China Mobile, China Unicom, ONF, ON.Lab, OPNFV, SK Telecom)
The demo will walk through the sequence of steps an operator would take to on-boarding a new service in to a running CORD deployment, using a third-party Virtual Truck Roll (VTR) service as an example. We will also pull back the covers to show how such a service is embedded into the collection of servers and switches that comprise the CORD POD, as well as how a Northbound Interface (and GUI) is auto-generated to support runtime calls on the service (e.g., to diagnose subscriber problems).