Multi-core I/O Trace Analysis Using Distributed Local Memory
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data storage systems face challenges in efficiently recording and aggregating trace information from multiple computing modules, leading to increased computational costs and delays due to the use of global memory, which is not locally managed and results in resource contention.
Innovation Solution
Processing cores record trace information in dedicated local memory and aggregate it across modules, reducing the need for global memory access by accessing and analyzing trace information from local memories of other computing modules, thereby minimizing communication costs and resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If trace information is recorded in global memory, then trace aggregation can be performed across multiple computing modules, but computational costs increase and delays occur due to resource contention
Solution Approach 1:
The patent segments the centralized global memory trace recording into distributed local memory recording across multiple computing modules. Each computing module maintains its own local memory for trace information, eliminating the bottleneck of centralized global memory access while enabling parallel trace aggregation operations.
Solution Approach 2:
The patent introduces a hierarchical memory dimension by adding local memory layers at each computing module level, while maintaining access to global memory only when necessary. This multi-level memory architecture transforms the single-dimension global memory access pattern into a multi-dimensional access pattern with local and global levels.
2Device complexity
If trace information is recorded in global memory, then centralized aggregation is simplified, but resource contention increases and computational efficiency decreases
Solution Approach 1:
The patent divides the monolithic global memory management system into segmented local memory management units distributed across computing modules. Each module independently manages its local trace memory, reducing contention and improving parallel processing efficiency while maintaining standardized access protocols.
Solution Approach 2:
Each computing module autonomously manages its own local memory for trace information without requiring centralized global memory allocation. The modules self-organize their local memory structures and independently perform trace aggregation, reducing dependency on centralized resources.
3Loss of energy
If trace information is accessed from local memories of multiple computing modules, then communication costs are reduced, but data access complexity increases
Solution Approach 1:
The patent segments trace data access into local-first access patterns, where each computing module primarily accesses its own local memory before consulting other modules. This segmentation reduces cross-module communication while maintaining the ability to access distributed trace information when needed.
Solution Approach 2:
The patent implements local quality optimization by prioritizing access to local memory resources at each computing module. The system is designed so that the majority of trace access operations can be satisfied locally without inter-module communication, with fallback mechanisms for accessing remote trace data only when necessary.
Data Source
AI summary
Improved mechanisms and techniques for recording and aggregating trace information from multiple computing modules of a storage system may be provided. On a storage system having multiple computing modules, where each computing module has multiple processing cores, processing cores may record trace information for I/O operations in dedicated local memory—i.e., memory in the same computing module as the processing core that is dedicated to the computing module. One of the processing cores may be configured to aggregate trace information from across multiple computing modules into its dedicated local memory by accessing trace information from the dedicated local memories of the other computing modules in addition to its own. The aggregated information in one dedicated local memory then may be analyzed for functionality and/or performance and additional action taken based on the analysis.