Erasure Code Placement Across High/Low Energy Storage Zones
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data storage systems inefficiently manage erasure codes, leading to sub-optimal energy conservation and increased power consumption due to uniform storage of redundant data without considering the complexity of encoding or the likelihood of data access, resulting in unnecessary energy expenditure.
Innovation Solution
Implementing a method to store erasure codes based on their likelihood of use and energy efficiency, where frequently accessed data is stored in low-energy zones and less frequently accessed data in high-energy zones, optimizing energy usage by varying storage locations within a storage medium based on priority and access patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If all redundancy data is stored uniformly in the same location, then data protection is simplified, but energy consumption increases due to unnecessary access of rarely used redundancy data
Solution Approach 1:
The patent segments redundancy data into different categories based on access probability and importance. High-priority redundancy data is stored in fast-access memory while low-priority redundancy data is stored in slower, lower-cost storage. This segmentation allows the system to avoid accessing rarely used redundancy data, reducing energy consumption while maintaining data protection capabilities.
Solution Approach 2:
The patent applies local quality by assigning different storage locations with different energy characteristics to different redundancy data based on their access patterns. Frequently accessed redundancy data is placed in high-energy, fast-access storage, while rarely accessed redundancy data is placed in low-energy, slow-access storage. This creates locally optimized storage quality that matches actual usage patterns.
2Reliability
If erasure codes with higher protection levels are used, then data reliability improves, but overhead and energy consumption increase
Solution Approach 1:
The patent implements dynamic redundancy management where the level of redundancy and its storage location can change over time based on observed access patterns. If redundancy data is rarely accessed, the system can reduce its protection level or move it to lower-energy storage. This dynamic adjustment allows the system to maintain high reliability when needed while reducing energy consumption during normal operation.
Solution Approach 2:
The patent changes the parameters of redundancy data storage by adjusting both the amount of redundancy and its physical location based on access probability. For data with low access probability, the system reduces redundancy parameters and stores in low-energy locations. For frequently accessed data, higher redundancy parameters are maintained in fast-access storage, optimizing the balance between reliability and energy consumption.
3Speed
If redundancy data is stored in fast-access memory, then data recovery speed improves, but energy consumption increases
Solution Approach 1:
The patent performs preliminary classification of redundancy data based on access probability before storage. By identifying which redundancy data is likely to be needed soon, the system pre-positions that data in fast-access memory while leaving other redundancy data in slower, lower-energy storage. This preliminary action ensures fast recovery speed for critical data without paying the energy cost for all redundancy data.
Data Source
AI summary
Methods and apparatus associated with storing data in high or low energy zones are described. Example apparatus include a data storage system (DSS) that protects a message using an erasure code (EC). A location in the DSS may have an energy efficiency rating or a latency. Example apparatus include logics that produce an EC that has a likelihood of use, that select a location to store the EC in the DSS based on the energy efficiency rating, the latency, or the likelihood of use, that store the EC in the location, and that compute an order of retrieval for an EC stored in the location. The order of retrieval may be based on the energy efficiency rating or the latency. The EC may also have a priority based on the number of erasures for which the EC corrects.