The way the world is connected is changing. Instead of running applications in far-flung data centers, they're increasingly being pushed to the edge of the network. And this, according to Ciena CTO Steve Alexander, is changing the way we view optics.

Alexander explained that with the rise of technologies, like edge compute and 5G networks, data centers have become far more distributed.

"We're gonna be able to take advantage of this general trend that the propagation distances for a lot of edge applications are shrinking," he said.

"If you look at what Google is doing with Stadia, what Amazon is doing with Outposts, what Microsoft is doing with Azure, they are pushing their clouds out to the edge of the infrastructure," he added. "We're trying to get to a data center that's a couple hundred kilometers away from us. We're not trying to get to a data center on the other side of the world, most of the time."

Shorter spans mean the optics can operate at higher line rates with fewer amplifiers. As a general rule, lower line rates can travel longer distances without signal degradation. And this is where the Ciena exec argues the bulk of innovation is going to take place moving forward.

"You've still got the long-haul business and all that, that's still here, but the thing that's improving people's cloud experience has been shortening that mean time to cloud," he explained. "As the distances come down that you have to go to get to the cloud, you can get rates up, you can get additional savings, and you can lower the cost per bit."

Alexander adds that higher performance optical networks and wider distribution of edge compute will be critical for 5G networks to realize their full potential.

"I don't think that everybody really understands that to get the true 5G experience ... you have to have the edge data centers," he said.

And with 5G, Alexander said applications will be distributed with the most latency-sensitive components running at the edge while other parts are run in the cloud data center. "That's the way you're going to get this whole thing to scale up," he said.

According to Alexander, it's not enough to build a faster network, it's essential to make it easily programable too.

"When we look at the opportunity around the optical side, it's all around getting a lot of programmable capacity deployed right out to the edge of the network," he said.

The importance of programable networks crystalized with the onset of the pandemic and the shift to remote work, Alexander said. "Probably the biggest learning that's come out of this COVID-19 thing, from the carrier side, has been you want bandwidth at the edge to be fungible. You gotta be able to exchange it between what I thought was going to be an enterprise service, what's going to be a residential service, what's going to be a wireless service," he said.

Alexander explained that operators that built parallel networks where there was a completely separate network for enterprise, residential, and wireless "they were in trouble" when the pandemic hit and traffic began shifting to the residential network en mass. However, network operators that had adopted what Alexander called "adaptive networks" were able to shift capacity to meet demand.

Moving forward, the Ciena CTO expects this programmability to be a differentiating factor.

"Even if [the optics] costs zero, if it takes three truck rolls and a PhD in optics you know that's going to be a problem," he said. "It's this whole concept of fungible. ... I have the ability to move it around and change it."