Anrin Chakraborti

PhD Candidate
Department of Computer Science
Stony Brook Univerity
Stony Brook, NY - 11790

DBLP
Linkedin
CV
Github
Personal


About Me

I am a fifth year PhD student at Stony Brook University, New York.
I am fortunate to be a part of the Network Security and Applied Cryptography (NSAC) lab with Radu Sion.
My work encompasses system security, systems, applied crypto, embedded hardware and clouds.

I am on the job market! [CV] [Research] [Teaching]

Education
PhD in Computer Science     Stony Brook University     2014 -
B.E in Computer Science and Engineering     Jadavpur University, India     2010-2014

Professional Experience
Research Intern     IBM Research (Thomas J. Watson Research Center)     May 2017 - August 2017
Research Intern     Cryptanalysis Lab, National Chiao Tung University, Taiwan     May 2013 - August 2013
Research Intern     Indian Institute of Technology, Kharagpur, India     May 2012 - August 2012

Teaching
CSE310     Computer Networks     2014 Fall
CSE101     Introduction to Computers     2015 Spring

Projects


ConcurORAM: Multi-client Concurrent ORAM
Oblivious RAM overheads are often unacceptable in multi-client scenarios as it is difficult to support parallel accesses without carefully re-designing the protocols. ConcurORAM is a multi-client ORAM that allows clients to query in parallel without assumptions of trusted third parties or inter-client communication. This not only reduces the assumption footprint but also results in better scalability.

Research: Anrin Chakraborti, Radu Sion, "ConcurORAM: High-Throughput Stateless Parallel Multi-Client Oblivious RAM", Network and Distributed System Security Symposium (NDSS) 2019 (acceptance rate: 17%, 89/521) [paper]



Oblivious RAMs for Range Queries
I worked on designing an efficient ORAM protocol specifically suited for range queries that reduces overall access latency by minimizing disk seeks. Efficient range queries are critical for several applications, and we show experimentally that locality-preserving ORAMs have significant performance advantages over standard protocols when deployed with file systems.

Research: Anrin Chakraborti, Adam Aviv, Seung Geol Choi, Travis Mayberry, Daniel Roche, Radu Sion, "rORAM: Efficient Range ORAM with O(log2N) Locality", Network and Distributed System Security Symposium (NDSS) 2019 (acceptance rate: 17%, 89/521) [paper] [code]



PD-DM: Plausibly-Deniable Device Mapper
I jointly worked on designing a device-mapper with plausible deniability for the linux kernel that also preserves locality of access by minimizing disk seeks. Our techniques are orders of magnitude faster than randomization based techniques for plausible deniability.

Research: Chen Chen, Anrin Chakraborti, Radu Sion, "PD-DM: An Efficient Locality-Preserving Block Device Mapper with Plausible Deniability", Privacy Enhancing Technologies Symposium (PETS) 2019 (acceptance rate: 22%) [paper]



dm-x: Protecting Volume Level Integrity
I worked on building an integrity-preserving device mapper for the linux kernel. A key advantage of dm-x is that data in the protected volume can be updated efficiently, which is not possible in other read-only systems.

Research: Anrin Chakraborti, Bhushan Jain, Jan Kasiak, Tao Zhang, Don Porter, Radu Sion, "dm-x: Protecting Volume-level Integrity for Cloud Volumes and Local Block Devices", ACM SIGOPS Asia-Pacific Workshop on Systems (APSys) 2017 [paper] [code]



DataLair
DataLair is a device-mapper for the linux kernel that provides plausible deniability for data stored in "hidden volumes" against powerful multi-snapshot adversaries. As a key component of the system, we built an efficient write-only oblivious RAM. Our solution is more secure than the (once) very popular "TrueCrypt" and more efficient that the state-of-the-art (at that time).

Research: Anrin Chakraborti, Chen Chen, Radu Sion, "DataLair: Efficient Block Storage with Plausible Deniability against Multi-Snapshot Adversaries", Privacy Enhancing Technologies Symposium (PETS) 2017 (acceptance rate = 22%) [paper] [follow-up]



History Independent Data Structures
History independent data structures are used in several application e.g. electronic voting machines. In this work, we analysed history independence from a security perspective and provided several important game-based definitions for designing secure history independent data structures. We also provide the design of a delete-agnostic file system.

Research: Sumeet Bajaj, Anrin Chakraborti, Radu Sion, "Practical Foundations of History Independence", IEEE Transactions on Information Forensics & Security (TIFS) 2015 [paper]



Ongoing Work


SqORAM: A Locality-Preserving Write-Only Oblivious RAM
I am building an efficient locality preserving write-only ORAM for the Linux kernel that is optimized for use with modern filesystems.

Research: Anrin Chakraborti, Radu Sion, "SqORAM: A Locality-Preserving Write-Only Oblivious RAM" [pre-print]



Concurrent Query Authentication
In this project, I jointly developed query authentication mechanisms based on memory checking and then integrated these mechanisms to support concurrency for authentication of multi-client databases.

Research: Sumeet Bajaj, Anrin Chakraborti, Radu Sion, "ConcurDB: Concurrent Query Authentication for Outsourced Databases"



Invisible Plausibly Deniable File System
I am working on a plausibly-deniably flash-based file system that can also plausibly deny its own existence leveraging special characteristics of NAND flash devices.

Research: Chen Chen, Anrin Chakraborti, Radu Sion, "INFUSE: Invisible Plausibly-deniable File System for NAND Flash"



Research Intern at IBM
As an intern with the secure systems group at IBM Research, I worked on a cloud-based mechanism to protect against malware in email attachments and web links, levaraging secure virtual machines. We published a technical report on this.

Research: Anrin Chakraborti, Rick Boivie, Zhongshu Gu, Mehmet Kayaalp, Ankita Lamba, Dimitrios Pendarakis, A Cloud-Based Service That Protects End-User Devices from Malware in Email Attachments and Web Links" (Work done as intern at IBM Research) [Technical Report]