Telecom IQ
Showing results for 
Search instead for 
Did you mean: 

Is there an economically optimal approach in migration to NFV?



Guest Author: Robert Haim, Principal Analyst, ACG Research 

 DZ_ACG_TelecomIQ_Banner.jpgCommunication service providers (CSPs) love minimizing their cost base while maximizing their revenue streams. Yes, I may have captioned the obvious here, but it is easy to see that these are conflicting goals as in, it takes money to make money! In recent years, SDN/NFV has been touted as the glue that converges these goals. As such, a wave of migrations started a few years ago, as CSPs began to virtualize their network infrastructures. Along the way, a chilling realization hit the CSPs: the migration is not as easy as advertised as they faced major hurdles in recruiting scarce (and expensive) labor and simultaneously finding ways to beat their competitors by faster migration period. CSPs were also told that SDN/NFV relieves them of vendor lock-in issues. But they soon realized that avoiding vendor lock-in comes with an increase in deployment complexity and costs. It really isn’t simple to integrate multiple best-of-breed discrete system components (e.g., hardware, NFVI, VIMs, Open vSwitches, SDN Controller, a host of VNFs and VNFMs). But if CSPs tried to use the other route and use a pre- or partially integrated solution, they faced the issues associated with vendor lock-in.


So, there is the dilemma for the CSPs. Is there a better way? The “happy-medium” or simply the optimal approach is to find a reliable partner who can help CSPs with solutions that are pre- or partially integrated and offer the flexibility to accommodate third-party products that the partner can integrate into its solution.


To determine the economic differences of having a reliable partner or going your own way (or do-it-yourself – DIY), a methodical approach is necessary. We need to conduct the economic analysis of these migration strategies. Each requires multiple steps that include integration, testing, and verification. In one approach, best-of-breed discrete products must be integrated (resulting in no vendor lock-in) and in the other approach, a field-proven, reliable partner is selected that can offer a base system and help with the full integration cycle with its own or, alternatively, using third-party products for those parts that the CSP may desire.  


Let’s look at a simplified NFV architecture stack.ACG_Figure1.png

Figure 1. Simplified NFV architecutre 


Let’s assume that the CSP decides to go the DIY route and attempts to integrate a set of best-of-breed products. Here are some problems that can arise from this approach:

  1. The CSP must hire its own task force and try to find and hire scarce and expensive labor.
  2. Using multiple vendors can lead to finger-pointing if an issue comes up, e.g., interoperability.
  3. Misaligned product roadmaps can complicate and increase the cost of life cycle management.
  4. Long testing cycles with treacherous road to completion. The following diagram covers the complexity involved in “testing” cycles of the solution components.



Figure 2.  Testing cycles of solution components


Now, let’s assume that the CSP can establish a strategic alliance with a field-proven, reliable partner. This partner has a pre- or partially integrated solution that can quickly be enhanced with either the partner’s own products or other third-party products. In this case, it would be easy to list the advantages:

  1. No need to worry about vendor lock-in since the partner agrees to integrate third-party products.
  2. TCO optimization with lower system roll-out, life-cycle management, and more predictable cost levels. ACG Research has proven this.
  3. Peace of mind, as risk is reduced by streamlining integration with a faster time-to-market.
  4. No finger pointing in this case as the partner is the focal contact point.


In conducting the analysis of the economic differences between a “best-of-breed” DIY approach versus a pre- or partially integrated solution, we must include labor costs, OPEX, and opportunity cost differentials. This must be done on a per VNF basis and then on a full end-to-end basis with all the different VNFs in play. ACG conducted this analysis and the results can be obtained from:


 Don't miss ACG Research Principal Analyst, Robert Haim, discuss this topic live at Mobile World Congress in the MWC Discussion Zone (booth #3E11). See the full MWC Discussion Zone agenda for more details on this and other sessions.



hpe.com_csp_nfv.png CSPResources.jpg NFVResources.png @hpe_csp.jpg @hpe_nfv.png



0 Kudos
About the Author


Read for dates
HPE at 2019 Technology Events
Learn about the technology events where Hewlett Packard Enterprise will have a presence in 2019.
Read more
Read for dates
HPE Webinars - 2019
Find out about this year's live broadcasts and on-demand webinars.
Read more
View all