Abhinav Narain

Address: 3337, Networking and Telecommunication Group
Klaus Advanced Computing Building
266 Ferst Dr.
Georgia Institute Of Technology
Atlanta, Georgia, USA
Phone: 404-408-1947
nabhinav3 - gatech . edu

I am a graduate student at the Department of Computer Science at Georgia Institute of Technology.
I work with Professor Nick Feamster. Currently, I am collaborating with Prof. Mung Chiang and Prof. Kyle Jamieson at Princeton University.

I am affiliated with NTG and Noise lab .
My area of Research is broadly in Computer Networks, Systems and Security.

I am currently trailing my adviser at Princeton University, and have an office in C-334, Dept. of Electrical Engg. (E-quad)
Current mailing address:
Abhinav Narain,
Computer Science,
Computer Science Building,
Princeton, NJ 08540
email : anarain@cs.princeton.edu

Academics

I received my Bachelors (Dual degree) in Computer Science and Engineering at the Dept. of Computer Science & Engineering , Indian Institute of Technology Madras.
I was advised by Professor V. Kamakoti. I have a minor in Physics.

Publications

Projects

I have worked in wireless networks, measurement in initial years of my Ph.D. My current research is broadly related to deniability, privacy and security on Power-line networks. The following image might give you a glimpse of power-line and power-line signature of various devices on it.

View of Power-line Channel
View of Power-line Channel

It is interesting to note that the noise generated by different devices can be modulated depending by its load. It gets more interesting as we can vary the load on computing devices by changing their workload and hence modulate the noise. An example modulation of the noise is shown below, with the spectrogram and decoded bits. Bit 1 corresponds to loop activity on the processor, while the bit 0 corresponds to idle processor. This is a physical layer covert channel on power-line network.


Spectrogram of modulated processor activity
Spectrogram of the modulated bits by laptop's processor activity
View of Power-line Channel
Decoded bits

I am working on two projects described below.

Powerline Whisperer

We have constructed a deniable channel on Power-line network for two parties plugged into the same power-socket. We model the adversary next to the transmitter/receiver on the same power-socket. We use Software defined Radios and a powerline coupler(front-end) for transmission and reception of covert bits on the channel.

Setup showing a Software Defined Radio with Powerline coupler front-end attached to laptop
Setup showing a Software Defined Radio with Powerline coupler front-end attached to laptop

Exfiltration using Powerlines

We are also interested in the amount and kind information leakage possible due to noise generated by the SMPS converter of computing devices like laptops or IoT devices. The following spectrograms show differences in the frequencies generated.
spectrogram of noise generated by UDP packets blast
Spectrogram of noise generated by UDP packets blast
spectrogram of noise generated due to computations on processor
Spectrogram of noise generated due to computations on processor

Our preliminary results show our ability to differentiate activities on a lenovo laptop.
ROC curve for differentiating different activities on a laptop

Talks

Travel Grants

Courses

Teaching Experience

Extra Curricular

I was the Vice President of Graduate Student Council(GSC) for the school of Computer Science at Georgia Tech (2014-2015), just before I left for Princeton.
I have been on several panels, some of them are:

Personal

Essentials

Other Interests

I read a lot of philosophy. I relate it to physics many a times which work almost right for my view of the world. A minor in Physics has given me a fascinating approach to understanding the world using Quantum Physics to Relativity. I play musical instrument in my spare time.