Management Server Non-Volatile Memory Wear Leveling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing server switching method does not account for the characteristics of non-volatile memory, leading to premature degradation when a specific server handles a high processing load, potentially shortening the lifespan of non-volatile memory and requiring premature replacement.
Innovation Solution
A management server that calculates an evaluation value for each server based on the cumulative write amount of its non-volatile memory and selects an optimal server to distribute the workload, thereby leveling the deterioration of non-volatile memories across multiple servers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If workload is concentrated on a specific server, then processing efficiency is improved, but non-volatile memory lifespan is reduced
Solution Approach 1:
The patent combines multiple servers into a server group that shares a common pool of non-volatile memory resources. The memory is virtualized and dynamically allocated across servers based on workload demands, allowing the system to consolidate processing efficiency while distributing memory wear across multiple devices. This merging approach enables load concentration benefits while preventing any single memory device from experiencing excessive wear.
Solution Approach 2:
The system dynamically changes memory allocation parameters based on workload conditions and memory health status. When a server experiences high workload, the system can reallocate memory resources from other servers in the group, or redirect future write operations to different memory devices. This parameter adjustment allows the system to maintain processing efficiency while actively managing memory lifespan through workload distribution.
2Speed
If non-volatile memory is used instead of HDD, then read/write speed is improved, but memory deterioration occurs with repetitive writes
Solution Approach 1:
The patent implements a feedback mechanism that continuously monitors the health status, wear level, and performance metrics of non-volatile memory devices. Based on this feedback, the system dynamically adjusts memory allocation and workload distribution. When a memory device shows signs of deterioration or reaches wear thresholds, the system redistributes workloads to healthier devices, thereby maintaining high-speed performance while extending overall system reliability and preventing premature failure.
Solution Approach 2:
The system employs dynamic memory management where memory device roles and allocations are not fixed but continuously adjusted based on real-time conditions. Memory devices can transition between active, standby, and replacement states, and workload routing is dynamically changed in response to performance and health metrics. This dynamic approach allows the system to fully exploit the high-speed benefits of non-volatile memory while actively managing durability concerns through adaptive resource allocation.
Data Source
AI summary
A management server includes: a memory; a communication interface coupled to a plurality of servers having a server processor and a non-volatile memory; and a management processor coupled to the memory and the communication interface. Data is written to the non-volatile memory by the server processor in accordance with startup of a program. The memory is configured to store a limit value of a cumulative write amount of the non-volatile memory. The management processor receives a cumulative write amount of the non-volatile memory from each of the plurality of servers and calculates, when a startup request for requesting startup of a first program is received, an evaluation value of the non-volatile memory in each server based on the limit value and the cumulative write amount, selects one startup server based on the evaluation values, and transmits an execution request for requesting startup of the first program to the startup server.