Fwd: Nikhil Vanjani on Oct 16

[Alan Karp]

Not a Security Lunch talk, but it is about access control.

--------------
Alan Karp

---------- Forwarded message ---------
From: Aditi Partap <adit...@stanford.edu>
Date: Mon, Oct 13, 2025 at 8:00 AM
Subject: Nikhil Vanjani on Oct 16
To: security...@lists.stanford.edu <security...@lists.stanford.edu>

Fully Adaptive Decentralized MA-ABE: Simplified, Optimized,

                       ASP Supported

                       Nikhil Vanjani

                 Thursday, October 16, 2025

                       Talk at 4:00pm

                      CoDA E201 & Zoom

Abstract:

Access control is a foundational problem in computer security and it

requires that only authorized users can access specific data.

Attribute-Based Encryption (ABE) addresses this challenge by binding

ciphertexts to access policies and decryption keys to user attributes.

However, traditional ABE relies on a single trusted authority holding a

master secret key: a central point of failure that limits real-world

deployment.  Multi-Authority ABE (MA-ABE) overcomes this limitation by

distributing trust among multiple authorities managing disjoint sets of

attributes.  But prior MA-ABE schemes either were proven secure in

presence of static corruption of authorities or relied on complex

"dual-subsystem" proof techniques that made them inefficient.

In this talk, I will present a streamlined security analysis showing --

perhaps surprisingly -- that the classic Lewko–Waters MA-ABE scheme

(EUROCRYPT 2011) already achieves full adaptive security, provided its

design is carefully reinterpreted and, more crucially, its security

proof is re-orchestrated to conclude with an information-theoretic

hybrid in place of the original target-group-based computational step.

By dispensing with dual subsystems and target-group-based assumptions,

we achieve a significantly simpler and tighter security proof along with

a more lightweight implementation.  Our construction reduces ciphertext

size by 33%, shrinks user secret keys by 66%, and requires 50% fewer

pairing operations during decryption -- all while retaining full

decentralization and collusion resistance.

If time permits, I will also briefly discuss how these proof techniques

extend to construct the first MA-ABE scheme for arithmetic span

programs, capturing a richer class of access policies.

This is joint work with Pratish Datta (NTT Research) and Junichi Tomida

(NTT Research).

Bio:

Nikhil Vanjani is a final-year Ph.D.  candidate at Carnegie Mellon

University, advised by Elaine Shi. His research lies at the intersection

of applied and theoretical cryptography, with a focus on developing new

functional encryption techniques and its applications to decentralized

systems.  He received his B.Tech from IIT Kanpur and M.S.  from CMU,

where his thesis "Multi-Input Inner Product Encryption: Function-Hiding

Instantiations without Random Oracles" won the department's Best Masters

Thesis Award.

--++**==--++**==--++**==--++**==--++**==--++**==--++**==
security-seminar mailing list
security...@lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/security-seminar