ETSI held the second NFV Plugtests in Sophia Antipolis January 2018. This is a good first step to help CSPs get more comfortable in deploying virtualization in their existing networks. The market adoption for NFV and SDN has been muted at best but we are beginning to see positive signs of something more than a virtualized entity in the network. However, I wonder are we building another silo? The NFV Plugtests ignores real world deployments where systems and the legacy network elements that will exist for many more years is excluded in the test methodology. Interoperability is mostly focused on the VNFs, MANO systems (Orchestrators), and the NFV platforms.
Do we really think any CSP is going to build a greenfield virtualized network and deploy completely new software stacks that don’t interoperate with the billions invested in legacy OSS/BSS? It reminds me of the post 1996 Telecom Act which resulted in a flood of VC funding leading to the creation of brand new CLECs and the emergence of backbone fiber optic carriers. Some of this was truly greenfield for both infrastructure and software suppliers. However, for the incumbents that still had billions invested in their management and operations stack the suppliers still had to get certified via the dreaded OSMINE process.
What ETSI needs to be thinking about is how to move future plugtests beyond the 3 principle areas of NFV platforms, MANO, and VNFs. ETSI needs to consider the legacy system environment which it ignored in the ETSI NFV MANO architecture. Legacy systems are not going away in the next 3 years much less the next decade.
- Patrick Kelly