About Me

I recently completed my PhD at Stanford University advised by Professor Nick McKeown and am currently a postdoctoral scholar at Stanford. Prior to my PhD, I received my bachelors degree in Engineering from Harvey Mudd College.

I am passionate about building high performance network systems by leveraging hardware/software co-design. I am currently working on the nanoPU, a novel CPU-NIC architecture designed to provide ultra low latency and high throughput. The nanoPU moves the entire network stack into line-rate hardware: programmable transport, core selection, and thread scheduling. It uses a novel register file network interface rather than the traditional DMA-based network interface in order to provide the lowest possible latency and most predictable performance. We developed an open source, end-to-end FPGA prototype that runs on AWS F1 instances using Firesim. The wire-to-wire latency through the nanoPU is just 69ns – an order of magnitude lower latency than state-of-the-art commercial NICs! The nanoPU also enables single-digit microsecond 99% tail latencies and message processing throughput of over 100 Mrps per-core. You can learn more about the nanoPU from our paper and this 30min talk.

Research Interests

Education

Publications

The nanoPU: Redesigning the CPU-Network Interface to Minimize RPC Tail Latency
Stephen Ibanez, Alex Mallery, Serhat Arslan, Theo Jepsen, Muhammad Shahbaz, Nick McKeown, and Changhoon Kim
To appear in USENIX OSDI 21
Open source artifact
30min Talk Recording


Zerializer: Towards Zero-Copy Serialization
Adam Wolnikowski, Stephen Ibanez, Jonathan Stone, Changhoon Kim, Rajit Manohar, Robert Soule
To appear in HotOS 2021


The Case for a Network Fast Path to the CPU
Stephen Ibanez, Muhammad Shahbaz, and Nick McKeown
HotNets 2019


Event-Driven Packet Processing
Stephen Ibanez, Gianni Antichi, Gordon Brebner, and Nick McKeown
HotNets 2019
Open source artifact


The P4-NetFPGA workflow for line-rate packet processing
Stephen Ibanez, Gordon Brebner, Nick McKeown, and Noa Zilberman
FPGA 2019
Open source artifact


s-PERC: A distributed algorithm to calculate max-min fair rates without per-flow state
Lavanya Jose, Stephen Ibanez, Mohammad Alizadeh, and Nick McKeown
SIGMETRICS 2019
Open source artifact


Teaching Experience

CS344: Build an Internet Router, Stanford University

P4 Language Tutorial Instructor

P4-NetFPGA Tutorial Instructor