Rhongho Jang
Wayne State University, MI·Computer Science·Early-career
Internet routers are built to forward data, not to understand it. When traffic needs real analysis, such as spotting an attack in progress, it is usually shipped off to distant cloud servers, which adds delay.
- Fields
Computer Science
Engineering
- Topics
in-network computing
programmable routers
GPU acceleration
- Grants
- 1 known grant, through Apr 2031
What they work on
- Develop memory management techniques that coordinate data transfer between routers and embedded graphics processing units (GPUs) so AI inference can run at high throughput inside the network core.
- Design a vision-inspired sketching method that converts network traffic into compact vectorized representations at line rate, the full speed of the link, to make downstream model processing efficient.
- Build generative models that operate on these traffic sketches and the integrated GPU hardware to detect complex and evolving network threats.
- Release open-source designs, tools, workloads and network datasets in a public repository so others can build deployable in-network intelligence systems beyond simulation.
More about the research
That delay is a problem for applications that must react immediately. A different approach is to give the network itself the ability to think, by placing powerful computing hardware where the data already flows.
Why it matters
Routers that analyze traffic as it passes could detect cyber threats in real time instead of waiting on remote cloud servers, benefiting network operators and the users they protect. The open-source designs, datasets and hardware configurations released by the project could let other researchers build and test deployable in-network AI systems rather than only simulating them.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Reza Rejaie
University of Oregon, OR·Computer Science
Design a multi-objective, approximation-based scheme that schedules monitoring queries on a single multi-stage programmable switch, balancing accuracy, latency and reporting load.
Dong Jin
University of Arkansas, AR·Computer Science
Build an in-network cyberattack detection system for power-grid communication networks that uses programmable data planes for high-speed, customizable packet inspection.
Divya Mahajan
Georgia Institute of Technology, GA·Computer Science·Early-career
Build a cross-platform framework that measures how thermal hotspots, frequency throttling, power spikes and energy limits restrict distributed AI training performance on GPUs and other accelerators.
The school
Wayne State University: 21 labs on PI Pages, $292M a year on research, 194 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.