We use cookies to make interactions with our websites and services easy and meaningful, and to better understand how they are used and to tailor your experience. You can read more about our cookie policy here. By continuing to use this site you are giving us your consent to do this.

November 17, 2017 by Alan DiCicco

Does OpenCORD have you tied up in knots?


Last week at the CORD Build 2017 conference, I presented an update on the Calix open source contributions and testing of our OpenCORD solutions. It was great to celebrate the shared success of the Open Networking Foundation (ONF), but it also left me wondering if the community is missing the big picture.

For those keeping track, Calix has made multiple contributions to the ONF vOLT-HA (virtual OLT-Hardware Abstraction) and ONOS controller projects. Within vOLT-HA, we helped develop the open source code for OLT/ONT software lifecycle management, as well as ONT auto registration. During solution testing this fall, we successfully demonstrated ONOS virtual control plan interoperability for IGMP Proxy, DHCP Relay, and 802.1x Network Access Control protocols. And we did it all twice! Once with integration partner Radisys and a second time with Ciena. It was exciting to share that AXOS and the E9-2 have proven integration with the OpenCORD reference implementation.

Yet I was stuck by how rigid the OpenCORD reference implementation is. Two hardware providers who spoke before me proudly made a point that they were a true SDN solution because their hardware ONLY worked with the OpenCORD implementation. The unsaid downside to those statements is that their hardware has zero value without the OpenCORD software, which everyone at the conference freely admitted will require several more years of development and field trials. Ugly.

Additionally, the hallway conversation made it clear that no service provider would fully adopt the OpenCORD reference implementation. CORD is vision and architecture, but OpenCORD is a very specific reference implementation. Each service provider was going to integrate only select portions of the reference design into their network. No big surprise. Every operator has a unique environment and will maximize the return on their CORD ‘architecture’ investments.

Put those two observations together and you can see why my presentation was so unique. Calix integrated the AXOS platform into OpenCORD, not the E9-2 system. Because the AXOS platform is a fully functional operating system, service providers can deploy any AXOS system today and migrate to OpenCORD in the future by simply shutting off modules within AXOS; and do so without any loss of functionality. Because AXOS is a modular operating system, it is also flexible enough to work with any CORD architecture, not specifically limited to OpenCORD. Calix was the only participant talking about migrating networks to a CORD future. AXOS fits in to OpenCORD, but stands out as the only practical solution available today.

Time to unravel the Gordian knot that is OpenCORD. Grab the rope end that matches your end state network. If you follow it back to the beginning, you’ll find AXOS