Get moving with interoperability recommendations, available resources and where to find results from NFV experts.

SANTA CLARA, Calif.– Working together is the key to moving forward.

That was one of the main takeaways for the capacity crowd at the pre-conference tutorial on network functions virtualization (NFV) interoperability at the recent Open Networking Summit (ONS). 

Experienced teams discussed achieving interoperability in an NFV world through collaboration among stakeholders, demonstrating the importance of interoperability for communications service providers (CSPs), NFV infrastructure (NFVI) and virtualized infrastructure manager (VIM) vendors, application (VNF) vendors and open-source projects alike. 

Ericsson’s Chris Price summed up how vendors build interoperable applications, total solutions and implement click-to-buy, plug-and-play, application portability, interoperable functions and end-to-end automation. Tutorial attendees walked away with recommendations from the experts, available resources and where to find results. You can get up to speed by following the 1.5-hour session below or checking out the slides.

The session started with an overview of the current landscape where the OpenStack cloud platform is the most frequently deployed infrastructure manager for NFV. Open source brings richness in choice along with challenges that can be addressed by ensuring all components support current OpenStack releases and adhere to and validate use of open APIs using the OpenStack-provided faithful implementation test suite (FITS). Although the interop guidelines and test suite were originally designed for public cloud interoperability they are valuable for complex solutions such as NFV as well.

Extensive standalone and full-solution testing provides vendors the ability to build offerings that are easy to integrate and perform well in telecom environments. In addition to OpenStack, Open Platform for NFV (OPNFV), the  ETSI NFV Industry Specification Group (ISG) and testing organizations such as the European Advanced Networking Test Center (EANTC), offer programs and test suites to validate all layers of an NFV solution.

EANTC provides independent, vendor-neutral testing expertise for innovative telecom technologies. Managing director Carsten Rossenhövel presented recommendations to evaluate assurance qualities along with composite comparisons for the single vendor, light multi-vendor and full multi-vendor solutions they’ve tested.

Rossenhövel also shared the results of 2016 VNF/NFVI interoperability testing commissioned by the not-for-profit New IP Agency. Results were very good with over two-thirds of combinations passing. All seven NFVI/VIM solutions tested were popular NFV-ready OpenStack distributions. Helping providers learn how their offerings perform in relevant NFV scenarios was a key test objective and subsequent testing showed significant improvements. Look for the results of the in-progress testing of MANO and VIM combinations in May 2017.

All EANTC testing and the first ETSI NFV Plugtest adhere to the draft ETSI NFV Release 2 specification and testing methodology TST007. The January 2017 Plugtest was conducted with NFVI/VIM, MANO and VNF vendors and open source project participation. Interoperability testing at the Plugtest ranged from 97.7 percent to 100 percent. As each round of test results become available, it is clear the participants incorporate the experiences into their offerings.

The OpenStack community also held a mini-Summit at ONS: OpenStack Networking Roadmap, Community and Collaboration featuring Neutron core and distinguished engineer from SUSE Armando Migliaccio, AT&T principal technical staff member Paul Carver and OpenStack Foundation ecosystem lead, Ildiko Váncsa. You can check out those slides here.

The importance of collaboration was threaded through the four-day event which showcased how leading CSPs and enterprises worldwide are moving fast thanks to the benefits of a virtualized network environment.

In their joint keynote, John Donovan, chief strategy officer and group president, AT&T Technology and Operations, and Andre Fuetsch, CTO, President AT&T Labs, said over 250,000 percent growth in network traffic growth in the past 10 years mandated an open, virtualized and automated approach. Fuetsch urged the audience to focus on open source, “a much faster vehicle.”

AT&T open sourced the Enhanced Control, Orchestration, Management & Policy (ECOMP) platform, which they’ve run in production on the OpenStack-based AT&T Integrated Cloud in production for over 2.5 years. In February, ECOMP merged with the Open Orchestrator project (OPEN-O) to form the Open Networking Automation Platform (ONAP).

A five-year plan is the industry norm for achieving NFV and software-defined networking (SDN). However, the challenges of moving to a virtualized environment can result in multiple release versions of components from multiple vendors, which increase the need for integration and interoperability testing.

Cover Photo // CC BY NC

  • You said “moving to a virtualized environment can result in multiple release versions of components from multiple vendors” — this is an area where experienced cloud architects can offer validated integration suggestions.