Nathan Farrington  
  • Nathan Farrington
  • Director of Software and System Architecture
    Rockley Photonics, Inc.
    766 E Colorado Blvd, Unit 104
    Pasadena, California 91101, USA

Bio

Nathan Farrington is the Director of Software and System Architecture at Rockley Photonics, where he leads the software engineering team, designs next-generation data center networking solutions, and bridges the gap between the data center networking and optical communications communities. Previously, Nathan was the founder and CEO of Packetcounter, a computer network software and services company. Before that he was a data center network engineer at Facebook developing the Open Compute Project top-of-rack switch (Wedge and FBOSS). Before grad school, Nathan worked for the US Navy in mobile robotics and situational awareness applications for the Department of Homeland Security.

Nathan graduated from the University of California, San Diego, with a PhD in Computer Science and Computer Engineering. He was advised by Amin Vahdat, now at Google, as well as George Porter, George Papen, and Yeshaiahu “Shaya" Fainman. Nathan’s dissertation topic was on novel optical communications for data center networks. Nathan has served on the TPC of OFC/NFOEC 2014 and 2015, and as a reviewer for numerous journals including IEEE/ACM Transactions on Networking, IEEE/OSA Journal of Lightwave Technology, IEEE Micro, and ACM SIGCOMM Computer Communications Review. He is a member of the IEEE, the OSA, and the ACM.

Recent Projects

2014

Invited Talk: The Dawn of Channelized Ethernet
Nathan Farrington
IEEE Sympoisum on High-Performance Interconnects (Hot Interconnects 22), Mountain View, California, USA, August 26-27, 2014 (Slides: PDF)

Building a data-center network with optimal performance and economy
Daniel Tardent and Nathan Farrington
Lightwave, July 10, 2014 (HTML)

Invited Talk: 100GbE Lambda Switching for Data Center Networks
Nathan Farrington
OIDA 100GbE per Lambda for Data Center Workshop, San Jose, California, USA, June 12-13, 2014 (Slides: Keynote 6) (Slides: HTML)

The OIDA and the Ethernet Alliance organized a workshop to study 100 Gigabit Ethernet over a single wavelength, for the data center market. In my opinion, it is much too early to be worried about this.

2013

Integrating Microsecond Circuit Switching into the Data Center
George Porter, Richard Strong, Nathan Farrington, Alex Forencich, Pang-Chen Sun, Tajana Rosing, Yeshaiahu Fainman, George Papen, and Amin Vahdat
ACM SIGCOMM, Hong Kong, China, August 2013 (PDF)

This is amazing stuff. You should read it.

Invited Paper: Facebook's Data Center Network Architecture
Nathan Farrington and Alexey Andreyev
IEEE Optical Interconnects Conference, Santa Fe, New Mexico, May 7, 2013 (PDF) (Slides: Keynote) (Slides: PDF)

Invited Paper: A 10 μs Hybrid Optical-Circuit/Electrical-Packet Network for Datacenters
Nathan Farrington, Alex Forencich, Pang-Chen Sun, Shaya Fainman, Joe Ford, Amin Vahdat, George Porter, and George Papen
OFC/NFOEC, Anaheim, California, March 2013 (PDF)

Invited Talk: How to Make Optical Communication Products that Facebook Will Want to Buy
Future Needs of "Scale-Out" Data Centers: An OIDA Workshop for Stakeholders
Anaheim, California, March 2013 (Slides)

2012

Optics in Data Center Network Architecture
Nathan Farrington
Dissertation, UC San Diego, December, 2012 (PDF)

Invited Talk at Microsoft Research
Redmond, Washington, October 2012

Hunting Mice with Microsecond Circuit Switches
Nathan Farrington, George Porter, Yeshaiahu Fainman, George Papen, and Amin Vahdat
ACM HotNets, Redmond, WA, October 2012 (PDF) (Slides: Keynote) (Slides: PDF) (Slides: MOV) (Slides: HTML) (BibTeX) (Project Website)

This paper proposes a new scheduling algorithm for fast, microsecond-scale data center circuit switches, called traffic matrix scheduling. An advantage of traffic matrix scheduling over the previous hotspot scheduling is that more traffic can be offloaded to the circuit-switched portion of the data center network, making circuit switches much more useful than previously thought.

A Demonstration of Ultra-Low-Latency Datacenter Optical Circuit Switching
Nathan Farrington, George Porter, Pang-chen Sun, Alex Forencich, Joseph Ford, Yeshaiahu Fainman, George Papen, and Amin Vahdat
ACM SIGCOMM Demo, Helsinki, Finland, August 2012 (PDF) (Project Website)

One weakness of the Helios paper is that the relatively long switching time of 30ms limited the usefulness of the resulting hybrid network. So we developed our own optical circuit switch that is 1,000x faster than Helios. We demonstrated our switch at the SIGCOMM conference.

Invited Talk at Broadcom
San Jose, California, May 2012

Invited Talk at Big Switch Networks
Palo Alto, California, May 2012

2011

Hardware Requirements for Optical Circuit Switched Data Center Networks
Nathan Farrington, Yeshaiahu Fainman, Hong Liu, George Papen, and Amin Vahdat
OFC/NFOEC, Los Angeles, March 2011 (PDF) (Slides) (BibTeX)

While working on the Helios project, we discovered that many of the optical communication devices commonly used in data center networks, such as electronic dispersion compensators (EDC) and transimpedence amplifiers (TIA), are not optimized for quickly recovering after a loss of light. However, we did find that such technology exists and future data center networks could be engineered for high performance when using optical circuit switches.

2010

Invited Talk at Amazon.com
Seattle, Washington, November 2010

Helios: A Hybrid Electrical/Optical Switch Architecture for Modular Data Centers
Nathan Farrington, George Porter, Sivasankar Radhakrishnan, Hamid Hajabdolali Bazzaz, Vikram Subramanya, Yeshaiahu Fainman, George Papen, and Amin Vahdat
ACM SIGCOMM, New Delhi, India, August 2010 (PDF) (Slides) (BibTeX)

Traditionally, local area networks, including data center networks, have been constructed using switched Ethernet. A typical 10G Ethernet switch uses 12.5W per port, and cost upwards of $500 per port or more. There is a cheaper source of bandwidth: optical circuit switching. This paper describes how we constructed a data center network using an optical circuit switch and what we had to do to achieve good performance.

Scale-Out Networking in the Data Center
Amin Vahdat, Mohammad Al-Fares, Nathan Farrington, Radhika Niranjan Mysore, George Porter, and Sivasankar Radhakrishnan
IEEE Micro, July/August 2010 (PDF) (BibTeX)

This is an overview paper of our recent work and a great place to start when learning about the field of data center networking. It covers: commoditization, merchant silicon, scaling Ethernet, and scheduling TCP flows.

Invited Talk at HP Labs
Palo Alto, California, March 2010

Report on WREN 2009 -- Workshop: Research on Enterprise Networking
Nathan Farrington, Nikhil Handigol, Christoph Mayer, Kok-Kiong Yap, and Jeffrey C. Mogul
ACM SIGCOMM Computer Communication Review (CCR), January 2010 (PDF) (BibTeX)

This is a collection of paper summaries and transcriptions from the first WREN workshop.

2009

Data Center Switch Architecture in the Age of Merchant Silicon
Nathan Farrington, Erik Rubow, and Amin Vahdat
IEEE Symposium on High-Performance Interconnects, New York, August 2009 (PDF) (Slides) (BibTeX)

One of the primary reasons for low server utilization in data centers is that the data center network is oversubscribed (underprovisioned). This is necessary due to the non-commodity (super-expensive) nature of current commercial data center switches. This paper describes how to construct a 3,456-port 10G Ethernet switch out of "merchant silicon", i.e. switch chips designed for low-end commodity Ethernet switches. It is likely that future data center networks will be constructed almost entirely from merchant silicon, just as modern data center servers contain commodity Intel or AMD processors.

Scaling Data Center Switches Using Commodity Silicon and Optics
Nathan Farrington, Erik Rubow, and Amin Vahdat
ACM SIGCOMM Poster, Barcelona, Spain, August 2009 (PDF)

A poster version of the Merchant Silicon paper.

PortLand: A Scalable Fault-Tolerant Layer 2 Data Center Network Fabric
Radhika Niranjan Mysore, Andreas Pamboris, Nathan Farrington, Nelson Huang, Pardis Miri, Sivasankar Radhakrishnan, Vikram Subramanya, and Amin Vahdat
ACM SIGCOMM, Barcelona, Spain, August 2009 (PDF) (BibTeX)

Both Layer 2 Ethernet forwarding (learning bridge, minimum spanning tree, broadcast) and Layer 3 IP routing (longest-prefix matching, routing protocols) have problems scaling to large numbers of servers in the data center. This paper describes a new forwarding mechanism called PortLand designed specifically for data centers with much better scaling properties than either traditional Layer 2 or Layer 3 forwarding. Now, at least from a network address viewpoint, it is possible to treat all hosts in a data center as interchangeable components. Additionally, a PortLand network will never contain a forwarding loop.

Multipath TCP under MASSIVE Packet Reordering
Nathan Farrington
UC San Diego Tech Report, June 2009 (PDF) (Slides)

This was my "Research Exam", an in-depth survey of a specific area leading to a Masters degree. I have always been fascinated by how simultaneously good and bad TCP is across a wide range of networks and communication patterns. One design feature of TCP is that duplicate acknowledgements are used as an early indicator of packet loss, and hence an early indicator of congestion. Unfortunately, packet reordering will cause the receiver to transmit duplicate acknowledgements, fooling the sender into thinking that the network in congested. This limits opportunities for parallelism in data center networks, which by design have multiple paths between hosts for both increased capacity and fault tolerance. This report looks at historical approaches for fixing TCP so that eventually we will be able to fully utilize multipath data center networks.

Before going to grad school, I worked for the Navy as a robotics research engineer.

2005

Transitioning Unmanned Ground Vehicle Research Technologies
Estrellina Pacis, H. R. Everett, Nathan Farrington, Greg Kogut, Brandon Sights, Ted Kramer, M. Thompson, David Bruemmer, Doug Few.
SPIE Proc. 5804: Unmanned Ground Vehicle Technology VII, Orlando, FL, March 2005

2004

Intelligent behaviors for a convoy of indoor mobile robots operating in unknown environments
Nathan Farrington, Hoa Nguyen, Narek Pezeshkian
SPIE Proc. 5609: Mobile Robots XVII, Philadelphia, PA, October 2004

Towards a Warfighter's Associate: Eliminating the Operator Control Unit
H. R. Everett, Estrellina Pacis, Greg Kogut, Nathan Farrington, S. Khurana.
SPIE Proc. 5609: Mobile Robots XVII, Philadelphia, PA, October 2004

Segway Robotic Mobility Platform
Hoa Nguyen, J. Morrell, K. Mullens, Aaron Burmeister, S. Miles, Nathan Farrington, K. Thomas, Doug Gage.
SPIE Proc. 5609: Mobile Robots XVII, Philadelphia, PA, October 2004

A Segway RMP-based robotic transport system
Hoa Nguyen, Greg Kogut, R. Barua, Aaron Burmeister, Narek Pezeshkian, D. Powell, Nathan Farrington, M. Wimmer, B. Cicchetto, C. Heng, V. Ramirez.
SPIE Proc. 5609: Mobile Robots XVII, Philadelphia, PA, October 2004

Maintaining Communication Link for Tactical Ground Robots
Hoa Nguyen, Nathan Farrington, Narek Pezeshkian.
AUVSI Unmanned Systems North America, Anaheim, CA August 2004

Enhancing Functionality and Autonomy in Man-Portable Robots
Estrellina Pacis, H. R. Everett, Nathan Farrington, David Bruemmer.
SPIE Proc. 5422: Unmanned Ground Vehicle Technology VI, Orlando, FL, April 2004

Maintaining Communication Link for a Robot Operating in a Hazardous Environment
Hoa Nguyen, Narek Pezeshkian, A. Gupta, Nathan Farrington.
American Nuclear Society 10th Intl. Conf. on Robotics and Remote Systems for Hazardous Environments, Gainesville, FL, March 2004