Cluster Operation Thread Parallelization for Passive Node Latency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing cluster systems face inefficiencies in performing operations like connect or disconnect operations due to sequential execution on passive nodes, leading to increased total time as the number of passive nodes in the cluster grows, making these operations time-consuming and inefficient.

Innovation Solution

A system and method that coordinates simultaneous operation requests across all passive nodes in a cluster, utilizing a storage system interface engine to generate threads and sub-threads for parallel execution, allowing for rapid operation completion by communicating requests through communication engines and initiator engines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If operations are performed sequentially on each passive node, then the operation can be completed reliably, but the total execution time increases significantly as the number of passive nodes increases

Engineering Contradiction:
Improveoperation completion reliabilityVSAvoidtotal operation execution time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the operation execution into independent threads, with each thread responsible for performing the operation on a specific passive node. This segmentation allows concurrent execution across multiple nodes while maintaining reliable completion tracking through thread management, resolving the contradiction between reliability and execution time.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic thread creation and management where the system automatically creates worker threads for passive nodes based on the cluster configuration. The thread pool dynamically adapts to the number of passive nodes, enabling flexible parallel execution that maintains reliability while minimizing total execution time.

Inventive Principle:
Principle #15Dynamics

2Device complexity

If operations are performed sequentially on each passive node, then the system structure remains simple, but the productivity decreases as the number of passive nodes increases

Engineering Contradiction:
Improveoperation execution structureVSAvoidoperation execution speed
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent merges multiple sequential operation executions into a single parallel execution framework. By combining the operation logic into a unified thread-based system that can execute across multiple nodes simultaneously, the patent maintains relatively simple system architecture while dramatically improving productivity through concurrency.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent creates a universal operation execution framework that can handle operations on any number of passive nodes through a standardized thread-based approach. This multi-functional design allows the same code structure to serve both single-node and multi-node scenarios, maintaining architectural simplicity while enabling high productivity in large clusters.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Adaptability or versatility

If the cluster has more passive nodes, then the cluster availability and resource pool increase, but the operation execution time increases due to sequential processing

Engineering Contradiction:
Improvecluster resource capacityVSAvoidoperation execution time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent segments the operation execution into independent parallel threads, one for each passive node. This segmentation enables the system to leverage the increased resource capacity of larger clusters without suffering from sequential processing penalties, as each node can be processed simultaneously while maintaining cluster availability benefits.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS8726072B1System and method for improving cluster performance using an operation thread for passive nodes
Publication Date: 2014.05.13 NETAPP INC
  • US8726072B1 patent drawing
  • US8726072B1 patent drawing
  • US8726072B1 patent drawing

AI summary

A system for providing improved cluster operation performance comprises a storage system and a cluster system communicatively coupled to the storage system. The cluster system comprises an active node and a plurality of passive nodes. The active node comprises a storage system interface engine and at least one initiator engine and each of the plurality of passive nodes comprises a storage system interface engine and at least one initiator engine. The storage system interface engine of the active node is configured to coordinate communication between the cluster system and the storage system, and simultaneously communicate an operation request from each of the plurality of passive nodes of the cluster system to the storage system.