Mirror, mirror, what lies ahead for network functions virtualization?

Posted by Mae Kowalke on Tuesday, July 12, 2016 with No comments

Although communications service providers (CSPs) are going ahead with deploying virtualized functions, they no longer view this as a way to reduce opex (at least, not in the short term), and acknowledge there’s quite a bit of upfront capex investment required, Light Reading Editor-at-Large Carol Wilson noted in a recent article, citing stats from Heavy Reading’s Spring 2016 Future of Virtualization Index.

Basically, while virtualization has clear benefits (especially in the long-term for competitive differentiation in a 5G environment), it’s more complicated to deploy than first expected, Wilson explained. This means CSPs are proceeding more cautiously and thoughtfully in their virtualization efforts, examining the business case for each function based on operator size, location, and existing services.

The result is that some CSPs are taking an early-adopter approach, while others are taking the wait-and-see road with slower-paced proofs-of-concept efforts. And, it looks like, for many CSPs, initially virtualization will be most lucrative for serving new areas or rolling out new services; operators are shying away from the expense and disruption of moving existing services to virtualization.

Below is a quick run-down of conclusions in the Heavy Reading index (based on data collected during May, 2016, from than three dozen—mostly Tier-1—CSPs worldwide).

Changing timetable - CSPs are making good progress identifying the functions they plan to virtualization by 2020, with most having done so for at least some of the functions. However, the timetable appears to be slipping for when those functions will be identified. It’s seeming more and more likely this first step won’t be complete during 2016.

Priorities - It’s pretty clear where virtualization priorities lie: network functions and customer premises equipment However, although more than half of CSPs have identified all or most of their high priority functions to virtualize, only a few have a complete timetable for testing and deployment. At the same time, while most CSPs are confident they’ll successfully get high-priority virtualization areas up and running by end of 2017, more than a quarter have none of their high-priority functions in production.

Deployment - Almost all CSPs have started development on high-priority functions they’re aiming to virtualize by end of 2017, with nearly a quarter of CSPs having done some work on half or more of those functions. It’s becoming more clear to CSPs what a realistic timeframe is for getting high-priority functions into production, and the conclusion is that it will take longer than expected: possibly not until 2019. Still, CSPs are becoming more confident that they’ll have completed their entire virtualization transformation by the end of 2020.

Investment - Also coming into sharper focus is the fact that virtualization will involve significant upfront investment, and it won’t pay off for another 3-5 years. CSPs expect spending on virtualization will increase this year and next, and that, while capex will likely begin to level off or decline by 2019 or sooner, it could take as long as 10 years for that to happen.


Categories: , , ,