Storage Server Recovery Sequence Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

As the number of bits stored in each memory cell increases in nonvolatile memory devices, the probability of data errors also increases, leading to performance reduction in storage server systems due to the need for frequent recovery operations.

Innovation Solution

A storage server system with a recovery manager that groups virtual machines based on workload characteristics, collects recovery information from storage devices, and generates a recovery sequence table to optimize recovery operations, providing an optimal recovery sequence to storage devices allocated to new virtual machines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If the number of bits stored in each memory cell is increased, then the storage capacity is improved, but the probability of data errors increases

Engineering Contradiction:
Improvestorage capacityVSAvoiddata error probability
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The system performs preliminary actions by pre-determining recovery sequences based on virtual machine workload characteristics and storage device attributes before errors occur. The recovery sequence table is generated in advance, so when errors do occur, the system can immediately execute the pre-planned recovery sequence without delay, thus maintaining high storage capacity while managing error risks proactively.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If recovery operations are performed frequently to guarantee data integrity, then data reliability is improved, but system performance deteriorates

Engineering Contradiction:
Improvedata integrityVSAvoidsystem performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system applies local quality by tailoring recovery sequences to specific local conditions - different virtual machines with different workload characteristics receive different recovery sequences. The system analyzes individual VM attributes (read-intensive, write-intensive, random access patterns) and assigns optimized recovery sequences accordingly, rather than applying a uniform recovery strategy across all storage devices, thus maintaining data integrity while minimizing performance impact.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system changes parameters by dynamically adjusting recovery sequences based on varying workload characteristics and storage device attributes. The recovery sequence table generator creates different recovery sequences for different scenarios (e.g., sequential access vs. random access workloads), and the allocator selects the appropriate sequence based on current system conditions, optimizing the balance between data integrity and performance.

Inventive Principle:
Principle #35Parameter changes

3Device complexity

If a uniform recovery sequence is applied to all storage devices, then device complexity is reduced, but adaptability to different workloads deteriorates

Engineering Contradiction:
Improverecovery sequence managementVSAvoidworkload adaptability
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The system segments the recovery sequence management into distinct functional components: a recovery sequence table generator that creates sequences based on workload characteristics, and a recovery sequence allocator that assigns them to specific storage devices. This segmentation allows complex adaptive recovery strategies to be implemented through modular components rather than a monolithic system, maintaining manageability while achieving high adaptability to different workloads.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4332773A1Storage server and operation method of storage server
Publication Date: 2024.03.06 SAMSUNG ELECTRONICS CO LTD
  • EP4332773A1 patent drawingFigure 1
  • EP4332773A1 patent drawingFigure 2A
  • EP4332773A1 patent drawingFigure 2B

AI summary

A method of operating a storage server which includes a plurality of storage devices and a storage node configured to control the plurality of storage devices is provided. The method includes grouping first virtual machines based on workload characteristics of the first virtual machines, to generate a virtual machine group table; receiving, from first storage devices, attribute information of the first storage devices, the first storage devices being allocated to the first virtual machines; collecting recovery information from the first storage devices; generating a recovery sequence table by determining recovery sequences based on the virtual machine group table, the attribute information, and the recovery information; and providing a corresponding recovery sequence to a second storage device of the plurality of storage devices, based on the recovery sequence table, the second storage device being allocated to a new second virtual machine different from the first virtual machines.