What is SDN?
The internet is built on a web of cables, wireless signals and software that connect billions of devices worldwide. Directing how information moves across that vast network is a complex and evolving challenge.
Traditionally, each device in the network has made its own decisions about where data should go, often using vendor-specific approaches. But as demand grows and more devices come online, this model makes network upgrades slow, costly and difficult — a problem often called "internet ossification."
SDN was developed to address this challenge. Instead of configuring each device separately to control how data flows, SDN shifts control to centralized software that manages traffic across the entire system. Think about SDN as an air traffic control system: Instead of each pilot deciding on a flight route on their own, a central control tower monitors the airspace and directs planes along the safest and fastest route in real time.
By separating decision-making software from the underlying hardware, SDN makes networks more flexible, easier to update and better able to adapt to new technologies and rising demand.
Building the architecture
NSF recognized that overcoming internet ossification would require rethinking how networks were designed and managed. Beginning in the early 2000s, the agency made sustained investments in research that laid the groundwork for SDN. Examples include:
Credit: Barrett Lyon / The Opte Project (CC By-NC 4.0)
Credit: Nicolle R. Fuller/U.S. National Science Foundation
A new way of networking
In 2003, NSF launched the 100x100 project to bring high-speed, more reliable 100Mb/s internet connections to 100 million U.S. homes. Achieving that goal required new ways to design, manage and secure networks at a massive scale.
The project led to OpenFlow, a protocol that allowed administrators to program networks from a centralized location. This simplified network management, optimizing network performance and resource use.
Infrastructure for innovation
In operation from 2006–2023, the NSF Global Environment for Network Innovation (NSF GENI) was a nationwide, virtual test bed that allowed researchers to develop and experiment with new ideas using real internet traffic.
NSF GENI supported early research on SDN deployments and facilitated technology transfer from universities to industry. This led many of the world's largest tech companies (including Google, Microsoft, Facebook, Yahoo, Verizon and Deutsche Telekom) to embrace SDN, making the internet faster, more secure and more accessible.
The internet of the future
NSF Future Internet Design and Future Internet Architecture programs, established in 2007 and 2010, respectively, funded research into fundamentally new internet architectures that could be prototyped and evaluated on the GENI test bed.
Leveraging SDN principles, researchers developed new capabilities for mobile devices, content delivery, user privacy and secure cloud computing — advancing approaches that were adopted by companies like Cisco and building a new generation of experts in large-scale internet design.
Expanding user control
Supported by the NSF Expeditions in Computing program, the Programmable Open Mobile Internet (POMI) produced widely used open-source tools for education and research, including Open Network Operating System, a distributed network controller, and Mininet, a network emulator for teaching and testing SDN systems before deployment.
POMI also supported early work on programmable forwarding planes, paving the way for the first fully user-programmable switch chips and the P4 programming language ("Programming Protocol-Independent Packet Processors"), which allowed network operators to customize and optimize network behavior.
Credit: POWDER Team, University of Utah
Lasting impact
NSF continues to invest in infrastructure, research and workforce development to drive innovation and advance network technologies.
The NSF Chameleon Cloud Testbed allows researchers to deploy SDN to rapidly develop new and enhanced cloud computing architectures, such as software-defined data centers and multi-cloud systems. Through the NSF Platform for Open Wireless Data-driven Experimental Research, researchers can build and test their own 5G networks, expanding SDN concepts into wireless and cellular systems. The NSF-led Artificial Intelligence Institute for Future Edge Networks and Distributed Intelligence, funded by NSF and the U.S. Department of Homeland Security, provides students with opportunities to work on real-world projects under the guidance of world-class experts in network systems and AI.
Such initiatives are driving workforce development and U.S. leadership in cutting-edge networking technologies, fueling economic growth, protecting national security and enhancing quality of life.