Virtual Disk Assignment for Compute-Only Nodes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional virtual computing systems face issues with network traffic congestion and performance degradation due to VM migration, as they fail to account for long-term network usage effects, leading to increased I/O latency and reduced throughput.
Innovation Solution
A method for classifying and assigning virtual disks based on characteristics to spread out input-output demand across compute-only and hyper-converged nodes, minimizing network usage during migration by identifying optimal node placements that reduce relative network usage and direct migration costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If VMs are migrated to minimize inter-VM communications over a network, then network performance between VMs is improved, but network traffic during migration increases causing network congestion
Solution Approach 1:
The system performs preliminary classification of virtual disks based on their characteristics (I/O patterns, size, access frequency) before migration. This allows the system to pre-determine optimal destination hosts that can accommodate specific VM types, thereby minimizing network traffic during migration while maintaining network performance. The classification and assignment happen in advance, reducing the need for network-intensive migration operations.
Solution Approach 2:
The system assigns different types of virtual disks to different hosts based on their specific characteristics. Compute-only nodes receive virtual disks suitable for their processing capabilities, while hyper-converged nodes receive virtual disks that can utilize their local storage resources. This local optimization reduces network traffic by keeping I/O operations local to the host that has the appropriate resources.
2Productivity
If virtual disks are assigned to compute-only nodes, then resource utilization is improved, but I/O latency increases due to network dependency
Solution Approach 1:
The system classifies virtual disks based on their I/O characteristics and assigns them appropriately. Virtual disks requiring low latency are assigned to hyper-converged nodes with local storage, while virtual disks that can tolerate higher latency are assigned to compute-only nodes. This selective assignment optimizes resource utilization while minimizing I/O latency for performance-critical workloads.
Solution Approach 2:
The system creates and manages virtual disk copies strategically. When a VM is migrated, the virtual disk can remain on the original host or be copied to the destination host depending on the VM type and I/O requirements. This copying approach allows compute-only nodes to access virtual disks over the network when necessary, while maintaining the ability to serve local I/O requests efficiently.
3Productivity
If multiple VMs are co-located on the same host, then resource efficiency is improved, but network card saturation occurs increasing network usage
Solution Approach 1:
The system classifies VMs based on their network I/O characteristics and assigns them to appropriate hosts. VMs with high network I/O requirements are distributed across different hosts to prevent network card saturation, while VMs with similar I/O patterns are co-located to maximize resource efficiency. The management processor monitors network usage and dynamically adjusts VM placement to balance resource efficiency with network capacity.
Data Source
AI summary
A system and method include classifying and assigning virtual disks accessed from compute only nodes. The method determines, by a management processor of a virtual computing system, characteristics for a plurality of virtual disks hosted on a plurality of hyper converged nodes in a cluster of nodes in the virtual computing system. The method further classifies, by the management processor, each of the plurality of virtual disks based on the determined characteristics and identifies, by the management processor, one of the plurality of virtual disks to host data for a virtual machine on a compute only node based on the classification to spread out input-output demand in the cluster, reducing probability of input-output bottlenecks and increasing cluster-wide storage throughput. The method also assigns, by the management processor, the identified virtual disk to host data for the virtual machine located on the compute only node.