Striped Volume Set Hotspot Reduction via Multi-Pattern Data Placement

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In storage systems, the use of identical data placement patterns for striping data containers across volumes leads to cascading hotspots, causing overload and reduced throughput as one volume becomes a bottleneck, affecting the entire striped volume set.

Innovation Solution

Implementing a technique that uses multiple data placement patterns to stripe data containers across volumes, allowing the Locate() function to compute the location of data container content, ensuring consistent content delivery and distributing load across nodes, thereby reducing hotspot formation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If identical data placement patterns are used for striping data containers across volumes, then the striping process is simple and consistent, but cascading hotspots occur causing overload and reduced throughput

Engineering Contradiction:

Improvesimplicity of striping processVSAvoidthroughput

Core Design Contradiction:

Ease of manufactureVSProductivity

Solution Approach 1:

The patent applies local quality by using different data placement patterns for different data containers within the same striped volume set. Instead of uniformly applying one pattern across all containers, the system selectively assigns patterns based on container characteristics, thereby distributing I/O requests across different volumes and preventing hotspots while maintaining striping functionality.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent changes the parameter of data placement pattern from a fixed identical pattern to a variable pattern selected from multiple available patterns. This parameter change allows the system to adapt the striping behavior to different data containers, distributing load more evenly across volumes and preventing cascading hotspots that would otherwise reduce throughput.

Inventive Principle:

Principle #35Parameter changes

2Stability of the object's composition

If identical data placement patterns are used for striping data containers across volumes, then the data placement is uniform and predictable, but bottlenecks form on specific volumes reducing overall system performance

Engineering Contradiction:

Improveuniformity of data placementVSAvoidsystem performance

Core Design Contradiction:

Stability of the object's compositionVSProductivity

Solution Approach 1:

The patent implements local quality by allowing different data containers to have different data placement patterns within the same striped volume set. This enables the system to maintain uniformity where needed while introducing variability in specific locations to prevent bottlenecks on particular volumes, thereby preserving overall system performance.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent introduces asymmetry by using non-uniform data placement patterns for different data containers. Instead of symmetric identical patterns across all containers, the system employs asymmetric pattern selection that distributes I/O requests more evenly across volumes, preventing bottlenecks and maintaining high system performance.

Inventive Principle:

Principle #4Asymmetry

3Productivity

If multiple data placement patterns are used to stripe data containers, then load is distributed evenly and hotspots are avoided, but the complexity of the striping process increases

Engineering Contradiction:

Improveload distribution efficiencyVSAvoidcomplexity of striping process

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent manages the complexity introduced by multiple data placement patterns through parameter changes - specifically by selecting patterns from a predefined set based on data container characteristics. This approach maintains productivity through even load distribution while controlling complexity through systematic parameter selection rather than arbitrary complexity.

Inventive Principle:

Principle #35Parameter changes

4Productivity

If multiple data placement patterns are used for striping, then bottlenecks are reduced and throughput is maintained, but the system requires more complex locate and computation functions

Engineering Contradiction:

ImprovethroughputVSAvoidcomplexity of locate function

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent balances throughput maintenance with locate function complexity by using parameter changes - selecting from multiple data placement patterns based on data container properties. The Locate() function is enhanced to handle multiple patterns, but this complexity is justified by the significant improvement in throughput and prevention of bottlenecks achieved through the multi-pattern approach.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS8880814B2Technique to avoid cascaded hot spotting

Publication Date: 2014.11.04 NETAPP INC

AI summary

The present invention overcomes the disadvantages of the prior art by providing a technique that stripes data containers across volumes of a striped volume set (SVS) using one of a plurality of different data placement patterns to thereby reduce the possibility of hotspots arising due to each data container using the same data placement pattern within the SVS. The technique is illustratively implemented by calculating a first index value, an intermediate index value and calculating a hash value of an inode associated with a data container to be accessed within the SVS. A final index value is calculated by multiplying the intermediate index value by the hash value, modulo the number of volumes of the SVS. Further, a Locate( ) function may be used to compute the location of data container content in the SVS to which a data access request is directed to ensure consistency of such content.