Network-Aware Virtualization Workload Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current container cluster management systems like Kubernetes are unaware of network topology and utilization, leading to inefficient workload placement and increased network latency, resulting in suboptimal data center efficiency and potential overprovisioning.
Innovation Solution
A method that involves a computer virtualization scheduler receiving network locality and utilization information to determine optimal virtualization workload placement, maximizing bandwidth and minimizing latency by prioritizing nodes based on network proximity and utilization, and integrating this awareness into Kubernetes and OpenShift scheduling systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If workload placement is performed without network topology awareness, then scheduling simplicity is maintained, but network latency increases and bandwidth utilization decreases
Solution Approach 1:
The patent introduces network topology information as an intermediary layer between the workload scheduler and the physical network infrastructure. This intermediary provides awareness of network paths, bandwidth, and latency characteristics without requiring the scheduler to directly manage complex network configurations, thus reducing latency while maintaining scheduling simplicity.
Solution Approach 2:
The system implements feedback mechanisms where network performance metrics (latency, bandwidth utilization) are continuously monitored and fed back to the workload placement algorithm. This feedback enables dynamic adjustment of workload placement decisions to optimize network performance without requiring manual intervention or complex manual configuration.
2Productivity
If workloads are randomly distributed across the data center, then resource utilization appears high, but actual network efficiency decreases due to suboptimal placement
Solution Approach 1:
The patent changes the parameters used for workload placement from simple resource availability metrics to network-aware parameters including topology distance, bandwidth capacity, and latency characteristics. This parameter transformation enables workloads to be placed based on network efficiency criteria rather than just resource availability, improving network productivity without requiring additional physical resources.
3Extent of automation
If manual node and pod labeling is used for proximity scheduling, then network awareness is achieved, but automation is lost and manual configuration is required
Solution Approach 1:
The system enables self-service automation where the workload scheduler automatically discovers network topology information and performs optimization without requiring manual labeling or configuration. The scheduler autonomously queries network infrastructure, processes topology data, and makes placement decisions based on network efficiency criteria, eliminating manual configuration efforts while maintaining full automation.
Data Source
AI summary
In some examples, a method includes: receiving, with a computer virtualization scheduler, network locality information for virtualization equipment; receiving, with the computer virtualization scheduler, network utilization information for virtualization equipment; and determining, with the computer virtualization scheduler and based on the received network locality information and the received network utilization information, virtualization workload placement in order to maximize available network bandwidth and minimize network latency between virtualization equipment.