True or False: Technology has become incrementally cheaper and better over the past 20 years?

True seems the obvious answer. However, while price and performance have improved in virtually every sector, one critical technology has not followed the curve. Despite its increased importance in the cloud-computing era, network infrastructure remains relatively expensive and has been slow to evolve.

“Network operations teams have become risk averse, because any time they changed anything the network could break,” said Dominic Wilde (pictured), chief executive officer of networking company SnapRoute Inc. “So they have had to live in a world of no.”

Wilde spoke with with Peter Burris (@plburris), host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, at theCUBE’s studio in Palo Alto, California. They discussed the problems impeding network innovation and why network costs remain high (see the full interview with transcript here). (* Disclosure below.)

Network is risk averse

Why has networking remained stagnant? “I think it’s because networking has not embraced or driven software economics, whereas compute has in many different aspects,” Wilde said.

Linux was the first trigger for disruption in the software market. Since then, constant innovation has created a market where prices decline yet performance levels rise. “You have an architectural innovation married together with an economic innovation at the software level,” Wilde added.

Yet while compute transformed, network sat on the sidelines. Used every day but ignored unless it broke, it was practical but unglamorous, the jeans and T-shirt to compute’s haute couture. Then software-defined networking brought network into the cloud-computing era. SDN revolutionized networking, making it agile, directly programmable, centrally managed, open standards-based, and vendor-neutral.

This meant centers could use white-box technology for network switchers and routing hardware. “There was great hope that this would drive a real economic revolution in networking,” Wilde explained.

In reality, SDN only increased the problems. “We added overlays over the top and abstracted the underlying network and added more layers of complexity and expense that added to the complexity and decreased manageability rather than helping it,” Wilde stated.

The need for separate network models, with different operational limits for on-premises, public and private cloud added costs and restricted options. This caused a bottle neck for companies wanting to adopt a hybrid-cloud model. Meanwhile, network operators struggled, attempting to work within a changing, dynamic cloud environment with “a set of tools and a set of products that only enable them to build a very static and very brittle, distributed … system, distributed network,” Wilde said.

A new paradigm in network operating systems

SnapRoute saw that cloud architecture could be harnessed to deliver fast, secure network service that was efficient and easy to manage. “We are in the container age … let’s embrace containers,” Wilde stated.

SnapRoute launched the industry’s first a cloud-native network operating system in February of this year. “It’s a fully containerized microservices architecture from the ground up,” Wilde said.

One benefit of the containerized approach is that it “looks like a set of programmable services to the DevOps world,” according to Burris. It also enables a new economic model without the “monolithic footprint” of the white-box approach.

Combining SnapRoute’s software with white-box infrastructure dramatically reduces overhead, according to Wilde. “We’re giving the opportunity to deliver control and the opportunity to innovate to operators. But, most significantly … we’re about 50% the price of any of the legacy, incumbent vendors,” Wilde said.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s CUBE Conversations.  (* Disclosure: SnapRoute Inc. sponsored this segment of theCUBE. Neither SnapRoute nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE