Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hiding Access-pattern is Not Enough! Veil: A Storage and Communication Efficient Volume-Hiding Algorithm

Shanshan Han, Vishal Chakraborty, Michael Goodrich, Sharad Mehrotra, Shantanu Sharma

TL;DR

Veil introduces a tunable, bucket-based volume-hiding scheme for encrypted key-value stores that randomizes key-to-bucket mappings and pads to equal bucket sizes, preventing adversaries from inferring query volumes. A direct version (Veil) uses random bucketing with a stash to control leakage and overhead, while Veil-O leverages a $d$-regular graph to enable overlapping buckets and reduce storage amplification. The authors provide analytic and experimental evidence showing Veil achieves near-optimal query amplification and low stash sizes, with Veil-O offering further storage savings at modest stash trade-offs. Combined, these methods yield scalable, dynamic-friendly volume-hiding that outperforms state-of-the-art approaches across skewed and large datasets, with practical setup and query performance advantages.

Abstract

This paper addresses volume leakage (i.e., leakage of the number of records in the answer set) when processing keyword queries in encrypted key-value (KV) datasets. Volume leakage, coupled with prior knowledge about data distribution and/or previously executed queries, can reveal both ciphertexts and current user queries. We develop a solution to prevent volume leakage, entitled Veil, that partitions the dataset by randomly mapping keys to a set of equi-sized buckets. Veil provides a tunable mechanism for data owners to explore a trade-off between storage and communication overheads. To make buckets indistinguishable to the adversary, Veil uses a novel padding strategy that allow buckets to overlap, reducing the need to add fake records. Both theoretical and experimental results show Veil to significantly outperform existing state-of-the-art.

Hiding Access-pattern is Not Enough! Veil: A Storage and Communication Efficient Volume-Hiding Algorithm

TL;DR

Veil introduces a tunable, bucket-based volume-hiding scheme for encrypted key-value stores that randomizes key-to-bucket mappings and pads to equal bucket sizes, preventing adversaries from inferring query volumes. A direct version (Veil) uses random bucketing with a stash to control leakage and overhead, while Veil-O leverages a -regular graph to enable overlapping buckets and reduce storage amplification. The authors provide analytic and experimental evidence showing Veil achieves near-optimal query amplification and low stash sizes, with Veil-O offering further storage savings at modest stash trade-offs. Combined, these methods yield scalable, dynamic-friendly volume-hiding that outperforms state-of-the-art approaches across skewed and large datasets, with practical setup and query performance advantages.

Abstract

This paper addresses volume leakage (i.e., leakage of the number of records in the answer set) when processing keyword queries in encrypted key-value (KV) datasets. Volume leakage, coupled with prior knowledge about data distribution and/or previously executed queries, can reveal both ciphertexts and current user queries. We develop a solution to prevent volume leakage, entitled Veil, that partitions the dataset by randomly mapping keys to a set of equi-sized buckets. Veil provides a tunable mechanism for data owners to explore a trade-off between storage and communication overheads. To make buckets indistinguishable to the adversary, Veil uses a novel padding strategy that allow buckets to overlap, reducing the need to add fake records. Both theoretical and experimental results show Veil to significantly outperform existing state-of-the-art.
Paper Structure (36 sections, 3 theorems, 2 equations, 14 figures, 3 tables, 7 algorithms)

This paper contains 36 sections, 3 theorems, 2 equations, 14 figures, 3 tables, 7 algorithms.

Table of Contents

  1. Introduction
  2. Settings
  3. Problem Definition
  4. Adversarial Model
  5. Overview of Veil
  6. Notation
  7. Components of Veil
  8. Bucket Creation $\mathcal{BC}$ ($\mathcal{D}$,$QA, SA, f$) $\rightarrow \langle$MAP, $\mathcal{B}, {\color{black}\mathit{stash}} \rangle$:
  9. Padding
  10. Data Outsourcing
  11. Query Evaluation $\mathcal{QE}(k_i) \rightarrow \mathcal{B}[k_i]$:
  12. Filtering:
  13. Veil with Random Bucket Creation
  14. Bucket Creation $\mathcal{BC}$
  15. Padding
  16. ...and 21 more sections

Key Result

theorem 1

Given a dataset $\mathcal{D}$, user-defiend $\mathit{SA}$ and $\mathit{QA}$, and fanout $f$, suppose $\mathcal{BC}$ creates a set of buckets $\mathcal{B}$ over $\mathcal{D}$ using $\textsf{MAP(*)}$. If $\mathit{SA} \geq 1$, $\mathit{QA} \geq 1$, and $f \geq 0$, then the average size of the stash, re

Figures (14)

  • Figure 1: An Intuitive Example of Unsecure Buckets
  • Figure 2: A 3-Regular Graph.
  • Figure 3: A 3-Regular Graph of Veil-O
  • Figure 4: Adjacency Matrix $\overline{\mathcal{M}}$
  • Figure 5: Impact of $f$ ($\mathit{SA}$=1.2, $\mathit{QA}$=1)**
  • ...and 9 more figures

Theorems & Definitions (9)

  • Definition 3.1: MAP
  • Definition 3.2: Well-Formed Buckets
  • Definition 5.1
  • Definition 5.2
  • Definition 5.3: Well-Formed Buckets with Overlap.
  • theorem 1
  • theorem 2
  • Definition B.1
  • theorem 3