Abstract:Increasing storage density exacerbates DRAM read disturbance, a circuit-level vulnerability exploited by system-level attacks. Unfortunately, existing defenses are either ineffective or prohibitively expensive. Efficient mitigation is critical to ensure robust (reliable, secure, and safe) execution in future DRAM-based systems. This dissertation tackles two problems: 1) protecting DRAM-based systems becomes more expensive as technology scaling increases read disturbance vulnerability, and 2) many existing solutions depend on proprietary knowledge of DRAM internals. First, we build a detailed understanding of DRAM read disturbance by rigorously characterizing off-the-shelf modern DRAM chips under varying 1) temperatures, 2) memory access patterns, 3) in-chip locations, and 4) voltage. Our novel observations demystify the implications of large DRAM read disturbance variation on future DRAM read disturbance attacks and solutions. Second, we propose new mechanisms that mitigate read disturbance bitflips efficiently and scalably by leveraging insights into DRAM chip design: 1) subarray-level parallelism and 2) variation in read disturbance across DRAM rows in off-the-shelf DRAM chips. Third, we propose a novel solution that mitigates DRAM read disturbance by selectively throttling unsafe memory accesses that might otherwise cause read disturbance bitflips without proprietary knowledge of DRAM chip internals. We demonstrate that it is possible to mitigate DRAM read disturbance efficiently and scalably with worsening DRAM read disturbance by 1) building a detailed understanding of DRAM read disturbance, 2) leveraging insights into DRAM chips, and 3) devising novel solutions that do not require proprietary knowledge of DRAM chip internals. Our experimental insights and solutions enable future works targeting robust memory systems.

What problem does this paper attempt to address?

The core problem that this paper attempts to solve is the DRAM (Dynamic Random - Access Memory) read disturbance problem. Specifically: 1. **As the technology node scales down, the vulnerability of DRAM chips to read disturbance increases**: Due to the significant reduction in DRAM cell size and the distance between cells caused by technology node scaling, when a DRAM chip accesses a row of cells, it may cause bitflips in the data bits of adjacent rows, which is the so - called "read disturbance". This phenomenon not only affects the reliability of DRAM but can also be exploited to break memory isolation, thus threatening the security and privacy of the system. 2. **Limitations of existing defense mechanisms**: Current defense measures against DRAM read disturbance are either ineffective or too costly to be widely applied. In addition, many existing solutions rely on obtaining proprietary DRAM circuit design information, which is not realistic for most systems. To solve the above problems, the author defines two research challenges: - **Protecting DRAM - based systems is becoming increasingly expensive**: As the technology node scales down, the vulnerability of DRAM chips to read disturbance intensifies, increasing the cost of protecting these systems. - **Lack of a universal solution**: Many existing read - disturbance mitigation solutions are only applicable to systems that can obtain proprietary DRAM internal structure information. To this end, the paper proposes research work in three aspects: 1. **In - depth understanding of DRAM read disturbance**: By strictly characterizing the read - disturbance vulnerability of modern commercial DRAM chips under different temperatures, memory access patterns, spatial characteristics, and voltage conditions, revealing the influence of these factors on read disturbance, and explaining their significance for future attacks and solutions. 2. **Proposing a new memory controller mechanism**: Using the understanding of the internal structure of DRAM chips and memory controllers, develop an efficient and scalable mechanism to reduce bitflips caused by read disturbance, and at the same time use sub - array - level parallelism and the variation of read disturbance between different rows to reduce the performance overhead of maintenance operations. 3. **A solution without proprietary knowledge**: Propose a new type of solution that does not require knowledge of the internal structure of proprietary DRAM chips, and effectively mitigate read disturbance by selectively restricting unsafe memory accesses that may cause read disturbance. In summary, this paper aims to achieve effective and scalable mitigation of DRAM read disturbance by establishing a detailed understanding of DRAM read disturbance, using insights from modern DRAM chips and memory controllers, and developing solutions without proprietary knowledge, ensuring that future DRAM - based systems can operate robustly.

Enabling Efficient and Scalable DRAM Read Disturbance Mitigation via New Experimental Insights into Modern DRAM Chips

Spatial Variation-Aware Read Disturbance Defenses: Experimental Analysis of Real DRAM Chips and Implications on Future Solutions

Understanding the Security Benefits and Overheads of Emerging Industry Solutions to DRAM Read Disturbance

An Experimental Characterization of Combined RowHammer and RowPress Read Disturbance in Modern DRAM Chips

Read Disturbance in High Bandwidth Memory: A Detailed Experimental Study on HBM2 DRAM Chips

Improving DRAM Performance, Reliability, and Security by Rigorously Understanding Intrinsic DRAM Operation

Self-Managing DRAM: A Low-Cost Framework for Enabling Autonomous and Efficient in-DRAM Operations

Improving DRAM Performance, Security, and Reliability by Understanding and Exploiting DRAM Timing Parameter Margins

RowPress: Amplifying Read Disturbance in Modern DRAM Chips

Sectored DRAM: A Practical Energy-Efficient and High-Performance Fine-Grained DRAM Architecture

Uncovering In-DRAM RowHammer Protection Mechanisms: A New Methodology, Custom RowHammer Patterns, and Implications

An Experimental Analysis of RowHammer in HBM2 DRAM Chips

Understanding and Exploiting Design-Induced Latency Variation in Modern DRAM Chips

Understanding the Interactions of Workloads and DRAM Types: A Comprehensive Experimental Study

RowPress Vulnerability in Modern DRAM Chips

Fundamentally Understanding and Solving RowHammer

Retrospective: RAIDR: Retention-Aware Intelligent DRAM Refresh

Improving DRAM Performance by Parallelizing Refreshes with Accesses

Reducing Performance Impact of DRAM Refresh by Parallelizing Refreshes with Accesses

Understanding and Improving the Latency of DRAM-Based Memory Systems

DRAMScope: Uncovering DRAM Microarchitecture and Characteristics by Issuing Memory Commands