SoK: Systematizing a Decade of Architectural RowHammer Defenses Through the Lens of Streaming Algorithms

2025-11-09

AI Summary

RowHammer (RH) threats continue to evolve, with decreasing activation thresholds impeding DRAM scaling, while fragmented defense research—stemming from inconsistent terminology, modeling approaches, and taxonomies—hinders systematic progress. This work reframes RH architectural defenses as a data-stream problem and introduces the first unified analytical framework based on streaming algorithms. We systematically survey 48 RH mitigation mechanisms published over the past decade and establish the first comprehensive taxonomy, characterizing each by design principle, security assumptions, and hardware compatibility. Innovatively, we integrate Reservoir Sampling theory with a novel StickySampling technique to derive the first mathematically provable safety bounds for RH mitigation. Our contributions include two practical implementation guidelines tailored to distinct hardware platforms, empirically demonstrating that the streaming perspective significantly improves defense efficiency, generality, and theoretical rigor.

Technology Category

• Data Mining & Knowledge Management: Data Stream Mining
• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

A decade after its academic introduction, RowHammer (RH) remains a moving target that continues to challenge both the industry and academia. With its potential to serve as a critical attack vector, the ever-decreasing RH threshold now threatens DRAM process technology scaling, with a superlinearly increasing cost of RH protection solutions. Due to their generality and relatively lower performance costs, architectural RH solutions are the first line of defense against RH. However, the field is fragmented with varying views of the problem, terminologies, and even threat models. In this paper, we systematize architectural RH defenses from the last decade through the lens of streaming algorithms. We provide a taxonomy that encompasses 48 different works. We map multiple architectural RH defenses to the classical streaming algorithms, which extends to multiple proposals that did not identify this link. We also provide two practitioner guides. The first guide analyzes which algorithm best fits a given RHTH, location, process technology, storage type, and mitigative action. The second guide encourages future research to consult existing algorithms when architecting RH defenses. We illustrate this by demonstrating how Reservoir-Sampling can improve related RH defenses, and also introduce StickySampling that can provide mathematical security that related studies do not guarantee.

Problem

• Systematizing architectural RowHammer defenses using streaming algorithms
• Analyzing protection trade-offs across different technologies and threat models
• Providing mathematical security guarantees for RowHammer mitigation techniques

Innovation

• Systematizing defenses via streaming algorithms
• Providing taxonomy for 48 architectural works
• Introducing StickySampling for mathematical security