eric.senunas

The Emergence of Software-Defined Networking Platforms - Are you thinking about SDN yet?

Discussion created by eric.senunas on Jul 10, 2014
Latest reply on Aug 29, 2016 by julie.gordon

By

 

Software-defined networking (SDN) market transition enables unprecedented value from our networks but while we are riding through the hype cycle for SDN, it has become obvious that one (but not the only one) of the inhibitors for widespread adoption is lack of standards and mindshare of a particular platform. An open, evolutionary approach that fuels innovation and differentiation is required.

 

Wait! Don’t we have Openflow? And vendor led solutions like Cisco ACI and lots of others? Sure we do, but while OpenFlow can be a critical component in an SDN architecture, it just defined the southbound interface so far (the ONF started a northbound API working group recently) and the various early controllers following this approach just supported OpenFlow on the southbound side, and due to lack of direction, all had their own application ecosystem: So applications were not portable between different controllers so customers are still locked in and innovation can’t happen at scale as the development community is fragmented in their support of  those controllers.

 

Furthermore did most of the initial Openflow-centric architectures largely expect greenfield approaches (despite some solutions in the market that support OpenFlow Hybrid Switching) and legacy equipment couldn’t be easily integrated. Outside of the data center, this is not easy to resolve, as legacy networks are pretty large, and as SDN is much more than network virtualization in the data center — this is also something that needs to be addressed. Vendor-led initiatives like the one mentioned before also keeps customers locked in and don’t embrace innovation through the community at scale.

 

But there is hope: the OpenDayLight Project provides a basis for a variety of open and standards based SDN platforms that can leverage the power of the community and include protocols like OpenFlow right away. As various vendors in the market will start to provide solutions on top of this project those applications written by the community become portable. Those vendors who will be able to gain mindshare in the community and developers for their platform through better support, more flexibility, better extensions will create a new center of gravity in the networking market.

 

The large incumbent vendors will resist change as they have the most to lose and so their efforts in the community will be halfhearted.  The platforms that succeed will be those that — besides having a critical mass to pull it off — capture mindshare and developers and which do effectively address the market expectation of an interoperable, easy to deploy SDN platform that is open and standards-based, that allows a best of breed approach with mix and match capabilities and which become relevant as they provide a rich partner ecosystem with ready-to deploy, turnkey applications.

 

We can expect that only a few of those platforms will emerge in the next two to three years before the SDN market crosses the chasm of having 10 percent of all switches in production SDN-enabled — according to various analysts.

 

Getting back to OpenDayLight: Another great advantage of the OpenDayLight architecture is clearly its flexibility. It is not tied exclusively to a single protocol but has the ability to support legacy, current, and future protocols as well as extensions to those — both northbound and southbound — without architectural changes. This assures investment protection and longevity of a solution. As as it provides a northbound API infrastructure that enables application portability the deployment risk for customers is greatly reduced as they will have choice.

 

The market transition is on its way. The coming years will determine who the winner(s) will be.

Outcomes