Three Key Issues with Integrating NFV/SDN and How Service Providers Can Solve Them

2017 is becoming the year of software-defined networking (SDN) and network function virtualization (NFV). Boosted by early adoption from carriers such as AT&T, communications service providers (CSPs) are beginning to ramp investments in the NFV and SDN ecosystems to the tune of $158 billion by 2021, according to research firm Technology Business Research. The move to a more virtualized, cloud-based network architecture is seen by some as the first layer of the Digital Transformation, where a more open, flexible architecture includes tools such as data analytics and automation.

The market has seen a series of ups and downs—no surprise given the expansive work at hand to virtualize networks built up over the course of a century—but now it seems to be moving more consistently. AT&T threw down the gauntlet to the rest of the industry with its announcement in February that it plans to virtualize 55 percent of its network to software by the end of 2017. The service provider noted at the time that it had beaten its internal goal for 2016, converting 34 percent of its network to SDN. The company’s 2020 goal of 75 percent suddenly seems achievable. AT&T also revealed its plans to allocate a significant portion of its capex budget to invest in integrated wireless and wireline solutions for its business customers, enhancing the reach of its software-driven solutions like SD-WAN.

AT&T is not the only provider making progress on virtualizing their networks. Internationally, Telefónica and China Mobile have also made significant progress on their SDN/NFV initiatives, taking different courses while being driven by the same need to drive down costs and increase efficiencies.  And according to a December 2016 virtualization index from research firm Heavy Reading, which “provides all CSPs with a compelling reality check for measuring their own progress in three critical areas: planning, deployment and spending,” 45 percent of providers surveyed plan to have at least 20 percent of their network virtualization efforts in live production by the end of 2017, and another 27 percent plan to have 50 percent or more in live production by year’s end.

The benefits of moving to a virtualized network infrastructure that providers like AT&T are finding so appealing are well documented, but worth repeating. The three main motivators include:

Reducing expenditures:

  • CAPEX savings through commodity network infrastructure
  • OPEX savings due to increased automation      

Improving operational efficiency and performance:

  • Centralized control improves traffic management and enables rapid failover in managing failures
  • Centralized traffic engineering provides a global view of the supply and demand of network resources
  • Implementation elasticity allows dynamic scaling of network and services

Accelerating time-to-market for new services

  • Features and capabilities can be added through software
  • Additional capacity can be added through configuration

Yet despite the early advances we’ve seen from service providers around the world, based on the timelines and projections above, providers are going to be living with hybrid networks—a combination of legacy networks and new virtualized networks—for quite some time, perhaps even decades. This brings a bevy of new challenges, especially on the OSS side of the house. The introduction of a software-based model into a predominately hardware-based network means new ways of managing those domains and the assets within them.