"This chapter of my journey has been the most defining. It embodies my core belief that research, when guided by vision and executed with rigor, can reshape industries and serve society. From academia to open-source communities, from government-funded programs to real-world systems, I’ve consistently focused on translating bold ideas into impactful platforms and vibrant ecosystems."
My time at Stanford was driven by a spirit of reinvention. As Executive Director of the Clean Slate Internet Design Program—with Nick McKeown as Faculty Director—I helped catalyze a rethinking of Internet architecture to meet emerging needs in security, mobility, and performance. This work laid the foundation for OpenFlow and software-defined networking (SDN), influencing academic research and industry roadmaps alike.
In parallel, I co-led the NSF-funded Expedition on Programmable Open Mobile Internet (POMI-2020). This cross-disciplinary, multi-year project brought together faculty and students across Stanford—including Paul Raj, Monica Lam, Bernd Girod, and Ramesh Johari—to explore programmable, open mobile architectures. The effort produced influential publications, early-stage prototypes, student-led startups, and over 40 Stanford PhDs.
I also co-founded and led the Stanford Platform Lab, a collaborative environment focused on datacenter-scale systems. There, I worked alongside John Ousterhout, Balaji Prabhakar, Mendel Rosenblum, Phil Levis, Christos Kozyrakis, and a generation of talented graduate students.
As SDN gained momentum, a visionary group—including Scott Shenker, Nick McKeown, Urs Hölzle, John Donovan, and others—formed the Open Networking Foundation (ONF), Open Networking Lab (ON.Lab), and Open Networking Summit (ONS). I was honored to be entrusted to lead ON.Lab and ONS, and eventually the merged ONF–ON.Lab organization.
Together, we built a global movement—standardizing SDN protocols like OpenFlow, launching open source platforms, and fostering deep collaboration among operators, vendors, researchers, and developers. At its peak, ONF’s community included hyperscalers (Google, Microsoft), global operators (AT&T, Verizon, NTT, DT, SKT), OEMs (Cisco, Juniper, Huawei, Arista), silicon providers (Broadcom, Barefoot), ODMs (EdgeCore, Quanta), and integrators (Radisys, Infosys).
Key open source platforms:
ONOS – Distributed SDN control plane
Trellis – Spine-leaf fabric for edge and core networks
SEBA / VOLTHA – Disaggregated broadband access architecture
Mininet – Lightweight emulator for SDN development and testing
P4 – Data plane programming language
CORD – Transforming central offices into agile cloud service hubs
Aether – Cloud-native 5G + edge infrastructure, including:
These platforms didn’t just push boundaries—they were deployed and adopted by real-world networks. Just as important, we built strong communities around each initiative to ensure sustainability and collaboration.
To accelerate awareness, adoption, and dialogue, I co-founded the Open Networking Summit (ONS). It quickly became the definitive venue for industry and academic exchange around SDN and open networking.
ONS milestones:
Keynotes by tech leaders like Urs Hölzle, Amin Vahdat, Albert Greenberg, Mark Russinovich
Operator vision shared by John Donovan, Andre Fuetsch, Ito-san, Elad Nafshi, and others
Thought leadership from Martin Casado, Nick McKeown, Scott Shenker, Vint Cerf
Entrepreneurial & VC insights from Vinod Khosla, Peter Levine, John Vrionis
Live, curated SDN demos from global projects and startups
ONS was eventually acquired by the Linux Foundation, evolving into the Open Networking & Edge Summit (ONES). It played a key role in scaling the SDN ecosystem and aligning industry-wide efforts.
ONF, ON.Lab, and ONS—along with partners like Stanford, Google, and AT&T—played a central role in the networking industry’s transformation to SDN. Leading this journey was one of the most rewarding chapters of my career.
In 2021, ONF spun out Ananki to commercialize Aether. Before completing its first funding round, Ananki and the ONF engineering team were acquired by Intel.
At Intel, I led engineering for the Edge AI Platform—a multi-tenant, cloud-managed orchestration layer for distributed edge infrastructure and AI workloads. The platform enabled zero-touch, zero-trust deployment and lifecycle management of edge AI applications across industrial, retail, and enterprise environments.
In 2024, Intel open-sourced the platform to empower the broader community:
When I joined NSF in 2003, the Internet research community faced a crossroads. Legacy architectures were limiting innovation, and funding programs needed a reset.
I helped launch two major initiatives:
FIND (Future Internet Design) – visionary research on new Internet architectures
GENI (Global Environment for Networking Innovations) – a national testbed for experimentation
Together with leaders like Scott Shenker, Larry Peterson, Jon Turner, Dave Clark, and others, we catalyzed the next wave of Internet research and helped fund early SDN and OpenFlow deployments in Internet2 and across campuses.
I also launched the Networking of Sensor Systems (NOSS) program with leaders Deborah Estrin, David Culler and others — seeding what would later become the IoT and Industry 4.0 domains.
I began my career at Washington University in 1987 starting as an Assistant Professor and rising to be a full professor, mentored by Jerry Cox and Jon Turner. Our work focused on:
High-performance gigabit switches using custom ASICs
Gigabit Low-latency networking interfaces using a custom ASIC
Large Internet routers built on cell fabric
A range of applications including multimedia conferencing and collaboration, electronic remote radiology, multimedia including movies on demand and more
We built and deployed gigabit networks across campus, and shared our technology via "Gigabit Kits" with other universities. This research led to Growth Networks, a Silicon Valley startup we co-founded that was later acquired by Cisco. Growth's architecture powered Cisco’s flagship Carrier Routing System (CRS) for many years.
In some ways, the industry’s current push toward ultra-low-latency, high-throughput AI infrastructure echoes our earlier efforts to bring gigabit networking to mainstream computing. That continuity—and challenge—excites me as I explore the infrastructure side of the AI era.