The Open Platform for Network Functions Virtualization* (OPNFV*) was launched in September 2014 as a consequence of the work of the European Telecommunications Standards Institute (ETSI) Industry Specification Group NFV (ETSI ISG NFV). The establishment of OPNFV resulted from the realization that the open source community could be catalyzed and harnessed to provide an open, integrated, carrier-grade reference platform, to deliver on the NFV vision, and accelerate development and adoption of NFV products and services. OPNFV is hosted by the Linux* Foundation.
The primary OPNFV project goals are to:
- Develop an integrated and tested open source platform that can be used to build NFV functionality, accelerating the introduction of new products and services.
- Contribute to and participate in relevant open source projects that will be leveraged in the OPNFV platform, and ensure consistency, performance, and interoperability among open source components.
- Establish an ecosystem for NFV solutions based on open standards and software.
- Promote OPNFV as the preferred open reference platform.
The OPNFV community currently contains over 60 members, including many the world’s leading network equipment providers, service providers, server OEMs and OS vendors. It is also starting to attract members from the academic world. Intel is a leading contributor to OPNFV, as well as a platinum member.
In contrast to traditional open source projects, which develop common components or platforms, OPNFV adopts an “Upstream First” methodology. OPNFV chooses and integrates relevant upstream open source projects and devotes its developer resources toward integration and testing of those upstream components, as well as closing functionality and performance gaps in them. OPNFV has been built on OpenStack*, supports multiple software-defined networking (SDN) controllers (e.g. OpenDaylight*, ONOS*, and OpenContrail*), and contains acceleration and deployment technologies such as Data Plane Development Kit (DPDK) and Fuel.
OPNFV projects range from engaging directly with upstream projects to internal OPNFV projects that address test or feature gaps in those upstream projects. Internal project proposals, priorities, and scope are proposed by the community and overseen by the OPNFV Technical Steering Committee (TSC). In no circumstances however, do these OPNFV internal projects result in a “fork” of any upstream projects. In the event the OPNFV community develops patches or contributions for particular upstream projects, those patches are always intended to be merged with the upstream project to prevent forking and fragmentation.
Since its formation in September 2014, OPNFV has grown considerably. The OPNFV developer community has grown to over 100 direct contributors with many more developers working on OPNFV-related projects for their own organizations.
In June 2015, only nine months after inception, OPNFV announced its initial release, code-named ARNO. The ARNO release provided an integration of key open source components such as OpenStack*, KVM*, OpenDaylight*, and Open vSwitch*. But just as importantly, ARNO also introduced a HW-agnostic development environment, a Continuous Integration (CI) infrastructure, as well as automated deployment and testing, tooling and documentation. The next release of OPNFV, code-named Brahmaputra, is scheduled for release in February 2016, and it will contain a much more richly integrated set of components and functionality. Intel is heavily involved in Brahmaputra and is contributing work associated with real-time KVM, Data Plane Development Kit-accelerated Open vSwitch*, service assurance metrics and capabilities, and a comprehensive test framework for vSwitch performance testing.
OPNFV is already finding its way into vendor and service provider labs for testing, and this process will accelerate with Brahmaputra and future releases. In May 2016, OPNFV will also host an inaugural Plugfest, where equipment vendors, ISVs, and service providers will come together to test and validate virtual network functions (VNFs) against the OPNFV platform. The event will be hosted in the United States by Cable Labs, in Louisville, Colorado.
Intel’s own SDN/NFV reference architecture, Intel® Open Network Platform (Intel® ONP), will be fully aligned with OPNFV. Intel ONP takes the latest OPNFV releases, blends it with Intel’s latest HW ingredients and focuses on NFV use case optimizations (performance, latency), such as Virtual Customer Premise Equipment (vCPE) and Virtual Evolved Packet Core (vEPC).
The future success of OPNFV will hinge upon not only the progress achieved in closing functionality and performance gaps in upstream projects, but also the availability and industry-wide adoption of open, standardized interfaces through which VNFs and Management and Orchestration entities can interface to the NFVI and VIM, and do so in an automated and interoperable manner. To make this possible, and to facilitate true "industrialization" of NFV, OPNFV will also institute a compliance program based on the standardization of OPNFV's test and characterization framework, which the industry can adhere to when consuming and commercializing OPNFV output.
NFV and SDN are challenging assumptions about how networks are designed, deployed, and managed and we are in the midst of a technological and business paradigm shift perhaps unprecedented in the industry. OPNFV has been realized to as an initiative to support, accelerate and standardize these momentous changes. Intel’s ambition is to be at the heart of that journey, supporting and collaborating with fellow community members and with the industry.
For more details, see www.opnfv.org and for developers, see also the OPNFV community wiki at https://wiki.opnfv.org/
About the Author
Mike Lynch is a senior Product Line Manager with Intel’s Network Platforms Group. He is responsible for business enablement of Intel’s solutions across Telco Core and Edge Infrastructure and also manages Intel’s software contributions into the OPNFV project
Contact Mike: michael.a.lynch@intel.com