Carl's epiphany: the box business was going to shrink; software was going to become king; the carriers' traditional revenue would stagnate or worse. More recently, I heard much the same things from AT&T. It's now the common wisdom. Ten years ago, I didn't put all that together and I think few others did.
To decide the right moves, he started with the question "What will Calix's customers need in ten years?" His first conclusion was that the carriers would need to find massive new revenues. That would require extreme flexibility, including a network designed to satisfy new demands rapidly. With revenues flat to down, the carriers would need to cut costs drastically to stay profitable. Inevitably, that would squeeze their suppliers.
Russo decided the answer was agile software that brought carrier products to market quickly and allowed using less expensive standard hardware.
Today, we call that SDN - Software Defined Networking - and AT&T has become a true believer. So have most of the biggest telcos. Carl's aim is right on target.
Calix's AXOS software is now controlling G.fast DSLAMs at dozens of their customers. Verizon is testing it on NG-PON2 fiber units. The demonstrations were impressive, but everything is still early.
AXOS has clearly defined North and South interfaces to work with just any controller and hardware. They are directly working with CORD and ECOMP in the lab. They plan an open ecosystem and are actively recruiting companies with related products. They bring an attractive market to partners, including being near exclusive suppliers to hundreds of telcos. They have units at both Verizon and AT&T, with good prospects for growing their market share at the bigger companies.
I'm not qualified to report in depth about SDN software. To this uneducated viewer, their products look very good. I'm sure Calix would be delighted to demonstrate for you. While best known in the U.S. and Canada, they have a recent win in Australia and an expanding presence in much of the world.