“P” Stands for Physical
Takeaway’s from the SDN & OpenFlow World Congress 2016
While preparing for a speech at Layer123’s recent SDN and OpenFlow World Congress, I had a staggering experience. I discovered the concept of the “P-CPE”. The “P” stands for, well, physical. It appears that our industry has reached the next level, where everything that really exists now requires a “P” in front of it. Who would have thought so 10 years ago?
As you can tell, this congress wasn’t a hardware event by any means. Hardware is becoming a commodity, and no one seems to be bothered - at least no one at this event. The industry appears to be facing a divide between the old-school HW oriented folks and the new world. The former did not turn up, so this blog is not an accurate reflection of today’s reality, but rather of the things yet to come. Also this congress was a Tier 1 loaded event, with only a few Tier 2/3 carriers visible. Does this mean that the Tier 2/3 carriers are all hardware people? Or, is it that they just didn’t have the time to work on new hypes and bubbles and chose to wait until the dust has settled and useful real-world solutions have materialized? I think probably the latter.
Universal CPE (uCPE) was everybody’s darling at the Congress, but only the respective VNFs and the way the whole system can be operated enable differentiation. Ease of use, operational simplicity and modern GUIs are key to success. Thus far, one of the golden business rules in our industry has been that you can’t win a project with a strong NMS, but you can lose it because of a weak one. In a world where hardware is becoming less relevant, this might not be valid any more. IT oriented customers will continue to strongly judge vendors on their ability to create a great look-and-feel, end-to-end system, the underlying network will likely still add value, but in a less influential way.
SD-WAN was all the rave, but maybe people were just happy that, finally, a useful, real-life application appeared out of the SDN/NFV hype. There was some talk about 5G, and how important SDN/NFV will be to enabling network slicing, but it was the wrong show for digging deeper. 5G seems to have replaced IoT as the mega driver everyone is referring to, even though there is still some time left until real 5G networks will be rolled out.
The combination of domain and parent controllers appears to be the new standard architecture, effectively replacing the good old EMS/NMS combo. The community had to acknowledge this, at least for carrier networks, as the one-controller-and-a-flat-network dogma has too many serious performance and scalability issues. Good news for the industry, so the issue left is: how to monetize this? No one believes that vendors will be able to make money on selling controllers, but there’s good reason to believe that adding functionality to controllers in the form of modules or apps will be a good way of generating revenues. So the combination of controller attached functions, like PCE or analytics, a slick MANO/VNF onboarding system, or a top-notch GUI, will likely become major differentiators going forward.
Finally, there were some interesting developments in an area which usually doesn’t make it to the headlines - protocols. Some service providers are concerned about the amount of protocol overhead in today’s networks. In fact, ETSI is leading the NGP initiative to substantially reduce protocol overhead. They are even contemplating removing the TCP from TCP/IP. It comes as no surprise, however, that incumbent router vendors are not among the key drivers, but it could be very interesting for companies that don’t have to protect an installed base or simply want to help their customers. Time will tell whether this is another storm in a teacup or early grumblings of a revolution.
If you want to learn more about making SDN and NFV deployments a reality for your company, download our free MercuryTM NFV Products brochure for more information.