Row Hammer Monitoring via Dual-Level Counter Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-density memory devices experience intermittent failures due to 'row hammer' issues, where repeated access to a single row causes data corruption in adjacent rows, especially in DDR3 systems with heavy workloads, leading to inefficiencies in monitoring and refreshing large memory capacities.
Innovation Solution
A system that monitors a subset of the most active memory rows, tracking access counts within a defined time window to detect row hammer events and triggers targeted refreshes on adjacent rows to prevent data corruption, using a memory controller to manage and refresh specific victim rows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If monitoring is performed on every row of memory individually, then data corruption detection accuracy is improved, but device complexity and monitoring cost become prohibitive
Solution Approach 1:
The memory monitoring system is segmented into two levels: (1) a first counter associated with each memory row that counts activations, and (2) a second counter that tracks the maximum value among all first counters. This segmentation allows the system to monitor every row for accuracy while using a simplified aggregation mechanism to reduce overall complexity and resource requirements.
2Reliability
If all memory rows are monitored and refreshed, then data integrity is improved, but productivity and system performance deteriorate due to excessive refresh overhead
Solution Approach 1:
The system applies local quality by differentiating monitoring and refresh intensity across different memory rows based on their actual access patterns. Rows with high activation counts (potential victim rows) receive targeted refresh operations, while rows with low activation counts use standard refresh protocols. This selective approach maintains data integrity for at-risk rows while minimizing unnecessary refresh overhead on other rows, thereby preserving system performance.
3Reliability
If refresh operations are performed frequently on all rows, then reliability against row hammer is improved, but energy consumption and time loss increase
Solution Approach 1:
The system implements partial action by performing enhanced refresh operations only on a subset of memory rows identified as victim rows through the counter monitoring mechanism. The second counter identifies rows where the first counter values exceed a threshold, indicating potential row hammer vulnerability. Refresh operations are then applied selectively to these identified rows rather than uniformly to all rows, reducing energy consumption and time loss while maintaining adequate protection against row hammer attacks.
Data Source
AI summary
A system monitors data accesses to specific rows of memory to determine if a row hammer condition exists. The system can monitor accessed rows of memory to determine if the number of accesses to any of the rows exceeds a threshold associated with risk of data corruption on a row of memory physically adjacent to the row with high access. Based on the monitoring, a memory controller can determine if the number of accesses to a row exceeds the threshold, and indicate address information for the row whose access count reaches the threshold.