Last year, the Open Network Forum (ONF) officially laid OpenFlow (OF) to rest. After years of work which knew ups and downs, the ONF finally decided to close the book on OpenFlow and move on.
I’ve traditionally been a protagonist for open, programmable networks. And yet it seems that there continues to be a big gap between theory and practice. I talked about this before in my previous white papers: SDN Great in Theory, Less in Practice and Don’t Give Up on OpenFlow Yet. In my humble opinion, SDN as it is implemented today is far from delivering on its promise of open, programmable networks.
So when the ONF announced the Stratum project, I eagerly read up and engaged in the conversation. But I find myself asking if this new initiative will indeed be a sequel of OpenFlow.
The answer, as per the ONF is an irrevocable no. According to the ONFs VP marketing, Tom Sloane, obviously feels that the new initiative (P4) will take OpenFlow to the next level: “P4 is poised to usher in the next wave: true programmability. The P4 language makes the chip layer programmable so you can do dynamic things to packets.”2
BUT Does P4 really have a better chance of success than its predecessor (OF)?
OpenFlow didn’t deliver and SDN is still far from realizing its potential
Using OpenFlow, SDN was supposed to deliver a programmable interface to the network, thereby enabling real-time network control and innovation. In reality, it achieved only a fraction of what was promised.
Fast-forward to March 2018, and the announcement that ONF and Google have joined forces to create the Stratum Project. Its stated mission: “to develop an open-source reference implementation for white box switches supporting next-generation SDN interfaces”.
So the question on everyone’s lips: is this SDN sequel going to be better than the original? Or are we witnessing the proverbial ‘pet cemetery’ scenario, where something’s been brought back to life that should have been left dead and buried?
Before we answer, let’s better understand what went wrong in the past, so we might be able to determine its likelihood of success this time round. Let’s take a trip down memory lane….
This white paper will analyze OpenFlow and P4 in terms of their potential of success. It will examine:
After we learn about the similarities and differences between OF and P4, we will offer a number of suggestions which may help us all ensure that we deliver on the promise of open, programmable networks.
With 5G around the corner, can we as an industry truly continue to wait for a programmable network? Isn’t it time we:
True network programmability is not easily achievable, but possible. To make it work, it needs to be in everyone’s best interest. And that depends on how we go about implementing change. As we’ve said before, a positive outcome requires participants to recognize the symbiotic relationship between all parties
Hayim joined ECI to lead its innovation center and spearhead the efforts in the areas of NFV and SDN. Hayim comes to ECI with much experience both in the SDN/NFV areas as well as in more traditional telecoms. Previous to joining ECI, Hayim held the position of Chief Cloud IaaS and SDN Architect at Toga Networks, Principal Architect at Tejas Networks, Founder and CTO of Ethos Networks and more.