Network Functions Virtualization (NFV) and software-defined networking (SDN) promise to save service providers millions in hardware costs and shave weeks off the time to deliver and scale network services. Unfortunately those are forward-looking advantages that have little impact on today’s project schedules or the time and effort it takes new services to go from testing to production-ready.
According to a recent survey conducted by Coleman-Parkes Research on behalf of Amdocs, getting the test-deploy cycle below nine weeks is the new benchmark! But it’s proving frustratingly elusive for engineering teams trying to balance intense market pressures to deliver broader coverage, faster performance and new services with the need to guarantee good quality of experience (QoE) for end users.
Aggressive delivery timelines are in direct conflict with the complexity of network architectures behind new IP-based offers, such as Voice over LTE (VoLTE), streaming HD video and augmented/virtual reality apps. Even with the right testing processes in place, the unknown impact of new technologies, traffic patterns, devices and interactions with legacy network elements make it difficult to keep projects on track and on budget.
New technologies, new requirements
Service providers are performing more testing every year. Over two-thirds of service providers surveyed indicated that their testing efforts have increased in 2016 and will again in 2017. Reasons cited for the rise in testing volume include customer demands to provide more services on a greater number of devices, increasing complexity of managing service quality across network domains – such as supporting seamless handoff for VoLTE services – and the ongoing need to optimize network performance to gain competitive advantage.
Unfortunately the increase in testing volume isn’t necessarily translating into faster test-deploy cycles, which is why many service providers are looking for ways to improve those ratios, without compromising service quality.
Managing multiple vendors, shifting timelines
The survey also revealed that many service providers are turning to third-party testing partners to augment engineering teams and provide added skills, domain knowledge and tools to reduce cycle times.
One strategy used by service providers is to follow the “single vendor” model, where a prime system integrator (SI) manages the overall end-to-end testing cycle. This approach is viewed as beneficial, particularly when technologies are relatively new or complex, such as VoLTE and NFV. Reduced costs, clear issue resolution, increased quality and faster delivery were highlighted as the main advantages of the prime SI approach. An alternative some service providers are using is to add domain specialists to conduct some of the testing or to assist with specific issues during crunch times.
The good news? Breaking the nine-week barrier is possible, but making it happen requires a well-defined project plan, knowledgeable partners to fill resource gaps, and the right tools to get it done right.
Interested in reading more? Download the research paper, “Network Testing Challenges and Approaches: Insights from Service Providers Worldwide.”