Skip to main content

Test-Bed Infrastructure Helps Community Development of OPNFV

By July 31, 2015November 28th, 2016Blog

Last November, we started a project—codenamed ‘Pharos’—to develop a test-bed framework for creating a federated NFV test capability, hosted by OPNFV member companies from across the community. Nine months later, we’re very excited about the participation so far, with 10 existing test-bed labs spanning North America, Europe and Asia and even more in the works planned by vendors, system integrators and service providers. Companies like Orange, who just opened the first two OPNFV test-bed labs in Europe, have created community access for evaluating real-world OPNFV environments. Ericsson, Intel, Huawei, China Mobile and Dell are also hosting labs which are helping to develop a federated OPNFV test-bed capability that is geographically and technically diverse.


To make this happen, we needed to develop a baseline specification for an OPNFV “compliant” test environment along with tools, processes and documentation to support integration, testing and collaborative development projects with needed infrastructure and tooling.Pharos provides a consistent view of test-bed infrastructures and their capabilities to facilitate developer/project access. Other artifacts include libraries of workloads, test-tools, best practices, etc. And in addition to the community-hosted labs, the Linux Foundation hosts an OPNFV infrastructure that is used primarily for Continuous Integration (CI), Build and Release Deployment/testing. After using the community labs for early development, the Linux Foundation infrastructure was used for deploying and testing the Arno release.


Why invest in a community test-bed lab?

These test-bed labs have become a valuable resource to our developers, providing a platform for early testing of integrated OPNFV ingredients within realistic NFV environments via an open, consistent, repeatable test domain. And while some developing can be done in virtual or more simplistic environments, many features require substantial resources that may not be available to many developers. These community-accessible test-beds allow for a range of scenarios across varying community needs.

Another advantage the labs provide is simplification– setup and evaluation of OPNFV technologies and platforms is a complicated process with significant a integration effort and a multitude of configuration decisions. Reference implementations running on a test-bed with well-documented hardware and software configurations and integration/configuration learnings help lower barriers for adoption by the community.

Additionally, Pharos enables us the community to test across different hardware platforms, architectures and environments to ensure that OPNFV really is a multi-hardware platform.



Moving forward, these community labs will be connected to CI infrastructure for continuous deployment and testing in diverse, multi-vendor environments. We’re developing common test-bed tools and an infrastructure manager for configuring a variety of test-bed deployments tools and methods.

Now that Arno has been released, the community is looking towards the next iteration, Brahmaputra. To make these labs even more resourceful, we’ll be developing automated deployment and testing capabilities, making improvements to documentation for developers, making specifications for virtual environments so developers can start experimenting without access to expensive infrastructures and expanding the scale of test environments. We’re also in the process of developing centralized reporting and anlytics with dashboards and a reservation system.

To learn more about the Pharos test-bed community labs project—and to stay up to date on the latest developments—visit the OPNFV wiki page.

About the author of this post

Trevor Cooper
Trevor CooperTrevor Cooper has over 20 years of experience developing and deploying new technology solutions
in Networking, Telecoms and Healthcare markets. He has lived in US, Canada, UK and South Africa (where he grew up). He has worked extensively with multi-disciplinary and dispersed teams developing mission critical hardware and software. His current role at Intel is Solutions Architect working on Software Defined Networking initiatives. He also leads an Open Source group in OPNFV and represents Intel in the ETSI NFV Test and Open Source Working Group.