Shared Data Queue for Virtual Machine Workload Balancing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In virtualization environments, the uneven processing times of virtual machines (VMs) on a physical machine lead to idle VMs waiting for long periods, resulting in underutilization of resources and reduced processing efficiency due to differing data processing complexities.

Innovation Solution

A data assignment method where data is stored in a single queue and dynamically assigned to idle VMs, allowing them to process data continuously, with the scheduler adjusting data sizes based on the processing capabilities of each VM to ensure balanced workload distribution.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If each VM has its own queue for data processing, then data processing independence is improved, but resource utilization deteriorates due to idle VMs waiting for data

Engineering Contradiction:

Improvedata processing independenceVSAvoidresource utilization

Core Design Contradiction:

Ease of operationVSProductivity

Solution Approach 1:

The patent merges multiple individual VM queues into a single shared queue that all VMs can access. This allows idle VMs to fetch data from the shared queue when their individual queues are empty, ensuring continuous processing and improved resource utilization while maintaining data processing independence through logical separation of tasks.

Inventive Principle:

Principle #5Merging (Combining)

2Device complexity

If data is assigned equally to all VMs, then workload distribution simplicity is improved, but processing efficiency deteriorates due to uneven task complexity

Engineering Contradiction:

Improveworkload distribution simplicityVSAvoidprocessing efficiency

Core Design Contradiction:

Device complexityVSProductivity

Solution Approach 1:

The patent implements dynamic workload distribution where the system continuously monitors VM processing speeds and adjusts data assignment accordingly. Faster VMs receive more data chunks while slower VMs receive fewer chunks, optimizing overall processing efficiency while maintaining relatively simple distribution logic through adaptive adjustment rather than complex pre-calculation.

Inventive Principle:

Principle #15Dynamics

3Adaptability or versatility

If VMs process data at different speeds, then task completion flexibility is improved, but overall processing time deteriorates due to idle waiting periods

Engineering Contradiction:

Improvetask completion flexibilityVSAvoididle waiting time

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent ensures continuous useful action by implementing a shared queue mechanism that allows VMs to continuously fetch and process data without idle waiting periods. When a VM completes its current data chunk, it can immediately fetch the next chunk from the shared queue, maintaining continuous processing and eliminating idle time caused by uneven processing speeds.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS9348632B2Data assignment and data scheduling for physical machines in a virtual machine environment

Publication Date: 2016.05.24 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

A data assignment method for a physical machine having multiple virtual machines (VMs) running thereon includes receiving data to be processed from a scheduler for assigning job data; storing the received data to be processed into a single queue; and in response to an idle VM being among the multiple VMs, assigning data in the queue to the idle VM to be processed by the idle VM.