Skip to main content

Open Source Software and the Network of Tomorrow

By November 6, 2017Blog

This post originally appeared on the DARPA Hackfest blog

When I look at innovative developments in software defined radio (SDR), I can’t help but view those in a larger context of how networking overall is being redefined in software. For years, networking was one corner of technology that seemed to defy the oft-repeated adage that software is eating the world. With its patchwork of arcane and proprietary network elements, complex standardization processes, and hardware-centric operations, networking simply seemed resistant to the changes affecting computation and storage.

What a difference a few years make! In short timeframe, we have seen the advent not only of SDR but also of Software Defined Networking (SDN), which decouples the control plane and data plane, and NFV (Network Functions Virtualization), which treats the capabilities that for years were built into monolithic hardware appliances as dynamic cloud applications. The hope is that this will allow operators to innovate much more quickly, and, in response to demand and subscriber usage, scale out and in. With 5G and the Internet of Things (IOT) on the horizon, most operators feel that cloud, SDN, NFV, and SDR frameworks will all be necessary to rolling out these services.

What I find even more extraordinary than the pace of innovation and change is how the breakneck pace is being driven: through open source resources. The operators have embraced open source software with gusto: no less venerable a company than AT&T has claimed that its network will be built on 80% open source components within 10 years. I’ve watched a very closed and competitive industry develop the sense of common community, a hallmark of the open source ethic. I’ve watched cutthroat competitors go out of their way to collaborate, both across company and geographic lines, and it’s been a great counterpoint to much of the larger political narratives around us.

I have seen great innovation begin to emerge from this collaboration: a start to bringing AI and machine learning to network ops and to building self-healing and self-scaling networks that automatically tune themselves to the user experience. On the other hand, the vast array of data and subscriber information means that security and privacy concerns become even more critical, and it’s unclear that traditional telecom expertise either within companies or with policy makers has prepared anyone for this impact. And as cool as these technologies are, the ongoing litany of abuses on social media platforms has taught us that depending on algorithms to run a platform on which millions of people rely has consequences.

What gives me hope is that doing this work in the open in fully transparent communities means that anyone can see the actual software being used and can comment on and change it. The network is being reimagined. Will that be for the eventual good of the consumer or not?

Heather Kirksey will be speaking at the DARPA Hackfest on Tuesday, November 14 at 18:45, at the NASA Ames Conference Center in Mountain View, Califronia.