Storage Bias Mode Switching for Cache-Coherent Data Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage devices supporting cache coherent interconnect protocols lack effective mechanisms to switch between host bias mode and device bias mode, leading to issues such as stale data copies and inefficient data access.
Innovation Solution
Implementing a storage device with a bias mode management system that uses a locality counter to track data access frequency, a snoop filter to manage cache coherence, and a mechanism to proactively switch bias modes based on access patterns, ensuring data freshness and efficient access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the storage device uses host bias mode for data access, then data freshness is maintained, but access latency increases due to host involvement in coherence management
Solution Approach 1:
The storage device dynamically switches between host bias mode and device bias mode based on access patterns and data freshness requirements. This dynamic mode switching allows the system to optimize between data freshness (host bias) and access speed (device bias) depending on the specific operational context, resolving the contradiction between reliability and time loss.
Solution Approach 2:
The system changes the bias mode parameter from static to dynamic, allowing transitions between host bias and device bias states. This parameter change enables the storage device to adapt its coherence management approach, selecting the optimal mode (host or device bias) based on real-time conditions, thereby balancing data freshness and access latency.
2Productivity
If the storage device switches between host bias mode and device bias mode, then access efficiency is improved, but system complexity increases due to mode management overhead
Solution Approach 1:
The storage device autonomously manages bias mode switching based on monitoring access patterns and determining data freshness status without requiring complex external coordination. This self-service approach to mode management improves access efficiency while minimizing the overhead complexity by making the system self-regulating.
Solution Approach 2:
The system implements feedback mechanisms to monitor access patterns and data freshness status, using this information to automatically determine when to switch between host bias and device bias modes. This feedback-driven approach optimizes access efficiency while keeping mode management complexity manageable through automated decision-making.
3Reliability
If the storage device maintains data freshness through host bias mode, then data coherence is ensured, but access speed decreases due to host processor involvement
Solution Approach 1:
The storage device dynamically adjusts its bias mode based on the specific data access requirements. For frequently accessed data, it switches to device bias mode for faster access, while for data requiring strict coherence, it maintains host bias mode. This dynamic adaptation resolves the contradiction between data coherence and access speed.
Solution Approach 2:
Different data regions or cache lines can be in different bias modes simultaneously, allowing the system to maintain host bias for data requiring coherence while using device bias for data where speed is prioritized. This local differentiation resolves the contradiction by applying the appropriate bias mode to specific data locations based on their access characteristics.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A storage device (120) is disclosed. The storage device (120) may include a storage (315) for a data and a controller (310) to manage access to the data in the storage (315). A mechanism (340) may automatically manage a bias mode for a chunk of the data in the storage (315), the bias mode including one of a host bias mode and a device bias mode.