After three days with some of the thought-leaders of NFV and SDN in Barcelona I will try to draw some conclusions, but you might call it random thoughts. If you agree, disagree or just want to add what I am missing, this post is open for comments.
- SDN is real, NFV is getting real very quickly.
- SDN is not yet in production at carriers; large OTT providers and large enterprise are running or starting implementations.
- SDN is being pushed by the vendors, NFV by the carriers.
- NFV brings the cloud to the proximity of the end-user.
- NFV is just a redefinition of Openstack-like technologies, with some specific requirements.
- The business-case for SDN at carriers is not yet defined. Cost reduction is not enough, new revenue generating services must be defined.
- With hardware becoming a commodity, new operating models between vendors and carriers are needed.
- Vendors need to differentiate on software and services, not on hardware.
- SDN is expected to bring automation possibilities.
- Network slicing is much easier with SDN than in todays networks.
- SDN and NFV can enable each other.
- Network overlays don’t bring you from A to B, they bring A to B.
- SDN is about decoupling the control-plane from the data-plane.
- Nobody is talking about the management-plane.
- Hybrid scenarios, with SDN and legacy networks, are going to be a challenge.
- Some vendors are still focusing on their own controller/API.
And then to summarize this summary; in 5 years time SDN will be unavoidable, in what form of shape I can’t yet predict. NFV will also be unavoidable, but I am not sure if it will still be called the same in 5 years time.