MapComp: A Secure View-based Collaborative Analytics Framework for Join-Group-Aggregation

TL;DR

MapComp tackles the efficiency bottlenecks of secure join-group-aggregation by introducing a view-based framework with a payload-independent secure materialized view. The key idea is to securely generate and maintain a join-alignment view using Alignment-PSI, decoupling the join keys from the payload so updates incur MPC-free refresh and remain compatible with plaintext GA protocols. This enables multiple GA strategies, including optimized sorting and bitmap-based approaches, to significantly accelerate JGA queries while preserving privacy under semi-honest two-party MPC. Empirical results show dramatic improvements, with up to 308.9x faster real-world queries and GA speedups up to 1140.5x over state-of-the-art baselines, highlighting MapComp’s practical potential for real-time secure analytics in dynamic data settings.

Abstract

Join-group-aggregation (JGA) queries are fundamental to data analytics, yet executing them collaboratively across different parties poses significant privacy risks. Secure multi-party computation (MPC) offers a cryptographic solution. However, existing MPC-based JGA approaches consider only a one-time query paradigm and suffer from significant performance bottlenecks. It executes expensive join operations from scratch across multiple queries and employs inefficient group-aggregation (GA) protocols, both of which hinder their practical use for scalable, real-time analysis. This paper introduces MapComp, a novel view-based framework to facilitate JGA queries for secure collaborative analytics. Through specially crafted materialized views for join and novel design of GA protocols, MapComp removes duplicate join workload and expedites subsequent GA, improving the efficiency of JGA query execution. To address the challenge of continuous data updates, our materialized view offers payload-independence feature and provides significant efficiency improvements in view refreshing with free MPC overhead. This feature, on the other hand, also allows further acceleration for GA, where we devise multiple novel protocols that outperform prior works. Notably, our work represents the first endeavor to expedite secure collaborative JGA queries using materialized views. Our rigorous experiments demonstrate a significant advantage of MapComp, achieving up to a 308.9x improvement in efficiency over the baseline in real-world query simulations. Moreover, our optimized GA protocols achieve up to a 1140.5x improvement compared to prior oblivious sorting-based solutions.

Figures (18)

• Figure 1: System architecture and workflow.
• Figure 2: A toy example of our materialized view, where $E = \langle E\rangle\xspace_0 \oplus \langle E\rangle\xspace_1 = (1, 0, 0, 1)$ is the intersection flag that indicates the first and fourth elements in $\textsf{X'}$ and $\textsf{Y'}$ (highlighted in green) are in the intersection. The vectors $\textsf{X'}$ and $\textsf{Y'}$ are derived by applying permutation $\pi_0$ and $\pi_1$ to the original join keys $\textsf{X}$ and $\textsf{Y}$, respectively. The flag vector $E$ is secret-shared between the two parties as $\langle E\rangle\xspace_0$ and $\langle E\rangle\xspace_1$.
• Figure 3: Functionality for Alignment-PSI $\mathtt{F}_{\mathtt{aPSI}}$.
• Figure 4: Alignment-PSI protocol $\mathtt{P}_{\mathtt{aPSI}}$.
• Figure 5: A running example of $\mathtt{P}_{\mathtt{aPSI}}$ where $X = (c, b, d, a), Y = (c, a, f), n_x = 4, n_y = 3$.

Theorems & Definitions (3)

• Definition 1
• Theorem 1
• proof