Storage System Request Management for LU Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage systems face inefficiencies during logic unit (LU) migration, leading to performance issues and backend overload due to the lack of dynamic adjustment in the number of individual requests for bulk requests, affecting host IO and resource utilization.
Innovation Solution
A method and device for managing a storage system that dynamically adjusts the number of outstanding individual requests based on the average response time length of completed requests, allowing more requests to be allocated to faster migration sessions and fewer to slower sessions, thereby optimizing resource use and reducing host IO impact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If as many free requests as possible are obtained from free request pools to process bulk requests during LU migration, then the migration speed is improved, but the host IO performance deteriorates and backend overload occurs
Solution Approach 1:
The patent implements dynamic adjustment of the number of requests for bulk operations based on real-time system conditions. The storage system monitors host IO performance and backend load, then dynamically modifies the request count parameter to balance migration speed and system stability. This transforms a static request allocation into a dynamic control mechanism that adapts to changing system states.
Solution Approach 2:
The patent introduces a feedback mechanism where the completion status and performance metrics of bulk requests are continuously monitored. Based on this feedback, the system adjusts the number of requests in subsequent bulk operations. The feedback loop includes tracking migration progress, host IO response times, and backend resource utilization to optimize request allocation.
2Productivity
If as many free requests as possible are obtained from free request pools to process bulk requests, then more requests can handle the bulk request, but resource utilization becomes inefficient with varying response times
Solution Approach 1:
The patent changes the parameter of request count from a fixed or maximum value to a dynamically adjusted value based on system conditions. By modifying this parameter according to backend load and migration progress, the system optimizes resource utilization. This prevents both over-provisioning (wasting resources) and under-provisioning (insufficient processing capacity).
Data Source
AI summary
Various techniques manage a storage system. Such techniques involve: in response to detecting that a first request of a plurality of requests initiated for a bulk request is completed, determining a response time length for the first request, the bulk request being used to migrate data from a first storage device to a second storage device, each request of the plurality of requests being used to read data from the first storage device and write data to the second storage device; determining an average response time length of the completed requests of the plurality of requests based at least in part on the response time length for the first request; and updating the number of the plurality of requests initiated for the bulk request based on the average response time length.