Dynamic Refresh Allocation for Row Hammer Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor memory devices, repeated access to certain memory cells can cause an increased rate of data degradation in neighboring cells due to the 'row hammer' effect, necessitating a balance between background and targeted refresh operations to prevent information loss.
Innovation Solution
A dynamic refresh operation system that monitors access patterns to determine the allocation between targeted and auto-refresh operations, using counters to adjust the proportion of each type of refresh operation based on detected 'row attacks' and their types, allowing for real-time adaptation of refresh strategies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If targeted refresh operations are increased to mitigate row hammer effect, then data reliability is improved, but refresh operation time consumption increases
Solution Approach 1:
The patent implements dynamic refresh allocation where the refresh controller adjusts the proportion of targeted refresh operations based on real-time detection of row hammer attacks. When attacks are detected, the system increases targeted refresh operations to protect victim rows, and when attacks are not detected, it reduces targeted refresh operations to minimize time consumption. This dynamic adjustment resolves the contradiction by making the refresh strategy adaptive rather than static.
Solution Approach 2:
The patent employs a feedback mechanism where the refresh controller continuously monitors access patterns to detect row hammer attacks and uses this information to adjust the refresh operation allocation. The system detects attacker rows through access pattern analysis and feeds this information back to modify the refresh strategy, increasing targeted refreshes for victim rows when attacks are detected and reducing them when not detected, thereby optimizing both reliability and time efficiency.
2Quantity of substance
If memory cell density is increased, then storage capacity is improved, but susceptibility to row hammer effect worsens
Solution Approach 1:
The patent applies preliminary anti-action by proactively detecting row hammer attacks through access pattern monitoring and preemptively increasing targeted refresh operations on victim rows before data degradation occurs. The refresh controller identifies attacker rows and their corresponding victim rows, then performs refresh operations on victim rows in advance to prevent data loss, thereby counteracting the harmful row hammer effect before it can cause damage.
Solution Approach 2:
The patent introduces an intermediary mechanism (the refresh controller with access pattern detection capability) that mediates between the high-density memory cells and the row hammer effect. This intermediary monitors access patterns, identifies attacks, and coordinates targeted refresh operations to protect vulnerable cells, thereby reducing the impact of row hammer effects in high-density memory configurations.
3Reliability
If background refresh operations are reduced to allocate time slots for targeted refresh operations, then row hammer protection is improved, but overall memory refresh efficiency worsens
Solution Approach 1:
The patent implements dynamic allocation of refresh operation time slots based on real-time detection of row hammer attacks. The system maintains a baseline of background refresh operations for overall memory health while dynamically allocating additional time slots for targeted refresh operations when attacks are detected. This dynamic time slot allocation resolves the contradiction by making the refresh schedule adaptive rather than static.
Solution Approach 2:
The patent applies partial action by selectively increasing targeted refresh operations only for victim rows affected by detected row hammer attacks, rather than uniformly increasing all refresh operations. The system identifies specific victim rows through access pattern analysis and applies refresh operations only to those rows, thereby providing row hammer protection while minimizing the impact on overall refresh efficiency.
Data Source
AI summary
Embodiments of the disclosure are drawn to apparatuses, systems, and methods for dynamic refresh allocation. Memories may be subject to row hammer attacks, where one or more wordlines are repeatedly accessed to cause data degradation in victim rows nearby to the hammered wordlines. A memory may perform background auto-refresh operations, and targeted refresh operations where victim wordlines are refreshed. The memory may monitor access patterns to the memory in order to dynamically allocate the number of targeted refresh operations and auto-refresh operations in a set of refresh operations based on if a hammer attack is occurring and the type of hammer attack which is occurring.