Mass Storage System Performance via Minimum Unit Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mass storage systems with multiple storage units face performance degradation due to the slowest unit determining data access time, and existing RAID solutions require excessive resources for data protection, leading to performance drops when operating at performance boundaries.
Innovation Solution
A method for reading and writing data in a redundancy protected mass storage system that uses a minimum number of storage units by determining available data, discarding or rebuilding data as necessary, and utilizing flags to manage buffer overruns, allowing the system to compensate for slower units and maintain performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If all storage units are accessed in parallel to meet performance requirements, then data rate is improved, but the slowest storage unit determines the overall access time causing performance degradation
Solution Approach 1:
The system issues read requests to multiple storage units in advance and determines which units have data available before actually reading. This preliminary determination allows the system to identify and utilize only the storage units that are ready, avoiding waiting for slower units that will not contribute to the current data transfer.
Solution Approach 2:
Instead of requiring all storage units to complete their operations, the system reads data from only the minimum necessary number of storage units that have data available. This partial action approach allows the system to achieve the required data rate without being constrained by the slowest unit, effectively using fewer units than the total available when sufficient data is present in a subset.
2Reliability
If redundancy protection is implemented using traditional RAID architecture, then data availability is improved, but the amount of extra storage capacity and performance required increases
Solution Approach 1:
The system dynamically changes operational parameters by determining, for each read request, which storage units have data available and adjusting the read operation accordingly. This parameter change allows the system to adapt to the current state of storage units, utilizing only those that are ready, thereby maintaining performance while implementing redundancy protection.
Solution Approach 2:
The system employs dynamic behavior by continuously evaluating which storage units have data available and adjusting the read operations in real-time. This dynamic approach allows the system to optimize performance based on the current state of each storage unit, rather than following a fixed access pattern that would be constrained by the slowest unit.
3Volume of moving object
If the minimum number of storage units is used to achieve required data rate, then system compactness is improved, but the system operates at performance boundary with no tolerance for slow units
Solution Approach 1:
The system performs preliminary determination of data availability in storage units before executing read operations. This advance checking allows the compact system to identify which of the minimum required storage units are ready to provide data, enabling the system to operate reliably at the performance boundary by avoiding dependencies on slow units that may not be ready.
Data Source
AI summary
The present invention relates to a redundancy protected mass storage system (1) with increased performance, and more specifically to a mass storage system (1) with multiple storage units (2). According to the invention, the resources that are essentially provided for compensating the damage of one or more storage units (2) are also used to enhance the system performance. For this purpose during reading or writing the storage system (1) just waits for the responses of a minimum number (n) of required storage units (2) to start reading or writing, respectively.