File Server Array Pipeline Transmission with Spot Instances

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Pipeline transmission in cloud environments faces delays due to unstable network conditions, particularly when using spot instances, which are unreliable and can be reclaimed suddenly, making it challenging to maintain stable data transfer among nodes.

Innovation Solution

The method involves building file server arrays with both stable and unstable nodes, where unstable nodes are grouped and linked in opposite directions to ensure timely data transfer, with each node receiving two copies of data and rebuilding the array in response to node status changes, allowing for cost savings through the use of spot instances.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If spot instances are used to reduce costs, then cost savings are achieved, but reliability deteriorates due to sudden reclamation

Engineering Contradiction:
Improvenode stabilityVSAvoidcost
Core Design Contradiction:
ReliabilityVSEase of manufacture

Solution Approach 1:

The system segments nodes into stable nodes (on-demand instances) and unstable nodes (spot instances), allowing differential treatment and management strategies for each type to optimize both reliability and cost

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system dynamically changes operational parameters such as data replication factors and transmission paths based on the stability status of nodes, adjusting behavior to maintain reliability while utilizing cost-effective spot instances

Inventive Principle:
Principle #35Parameter changes

2Speed

If pipeline transmission is used for data transfer, then transmission speed is improved, but delays occur due to unstable network conditions

Engineering Contradiction:
Improvedata transfer speedVSAvoidtransmission stability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The system performs preliminary actions by pre-establishing multiple transmission paths and pre-replicating data across stable nodes before transmission begins, ensuring that alternative routes are ready if instability occurs during pipeline transmission

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system adds dimensional redundancy by creating parallel transmission paths through different node sequences, transforming a single-path transmission problem into a multi-dimensional path selection problem that can withstand node failures

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Productivity

If unstable nodes are included in the array, then cost efficiency is improved, but network stability deteriorates

Engineering Contradiction:
Improvecost efficiencyVSAvoidnetwork stability
Core Design Contradiction:
ProductivityVSStability of the object's composition

Solution Approach 1:

Stable nodes act as intermediary buffers between data sources and unstable nodes, absorbing the instability of spot instances while maintaining overall system reliability through their guaranteed availability

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system creates multiple copies of data and distributes them across both stable and unstable nodes, ensuring that even if unstable nodes are reclaimed, data integrity is maintained through redundant copies on stable nodes

Inventive Principle:
Principle #26Copying

Data Source

PatentUS11792066B2File server array and enhanced pipeline transmission
Publication Date: 2023.10.17 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US11792066B2 patent drawing
  • US11792066B2 patent drawing
  • US11792066B2 patent drawing

AI summary

In an approach for building file server arrays with stable and unstable nodes for enhanced pipeline transmission, a processor builds an array from a plurality of stable nodes, wherein each stable node of the plurality of stable nodes is linked to two other stable nodes of the plurality of stable nodes forming a line. A processor divides a plurality of unstable nodes into one or more groups of unstable nodes. A processor links each group of unstable nodes to two neighboring stable nodes within the array. A processor sends data through the array and the one or more groups of unstable nodes in two opposite directions. A processor monitors a node status for each node of the plurality of stable nodes and the plurality of unstable nodes.