Service VM for Multi-Hypervisor Storage Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current virtualization environments face inefficiencies in managing I/O and storage devices due to coarse-grained administration and limitations in storage-related optimizations, particularly in shared storage setups where multiple virtual machines share resources, leading to suboptimal resource utilization and performance.

Innovation Solution

The implementation of a Service VM that manages storage devices by virtualizing all hardware as a global resource pool, allowing direct storage and I/O optimizations within the primary data path, and utilizing IP-based requests to coordinate storage across multiple nodes, enabling finer-grained administration and improved performance through deduplication, QoS, and other optimizations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple virtual machines share storage resources in a traditional virtualization environment, then resource utilization improves, but storage administration granularity becomes coarse and performance optimizations are limited

Engineering Contradiction:
Improveresource utilizationVSAvoidstorage administration granularity
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The patent segments storage administration into two distinct layers: a Service VM that handles fine-grained storage operations (deduplication, snapshots, QoS) and hypervisors that manage VM lifecycle. This segmentation allows storage tasks to be administered at the individual virtual disk level rather than at the coarse volume level, resolving the contradiction between resource sharing and administration granularity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The Service VM acts as an intermediary layer between the hypervisors and storage devices. It receives storage operation requests from multiple hypervisors, performs fine-grained storage optimizations, and manages shared storage resources. This intermediary enables both efficient resource utilization and granular storage administration simultaneously.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If a single hypervisor type is used in the virtualization environment, then system simplicity is maintained, but adaptability to different virtualization platforms is reduced

Engineering Contradiction:
Improvesystem simplicityVSAvoidhypervisor type compatibility
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The Service VM is designed with universal functionality to work with multiple hypervisor types (VMware, Hyper-V, KVM, Xen). It provides a standardized interface for storage operations that can adapt to different hypervisor protocols and architectures, enabling a single Service VM design to serve diverse virtualization platforms without compromising system simplicity.

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

3Productivity

If storage optimizations are implemented at the hypervisor level, then performance improves, but compatibility across different hypervisor types becomes difficult to maintain

Engineering Contradiction:
Improvestorage performanceVSAvoidhypervisor type compatibility
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

By positioning the Service VM as an intermediary between hypervisors and storage devices, the patent enables storage optimizations to be implemented in a hypervisor-agnostic manner. The Service VM translates various hypervisor-specific storage requests into unified storage operations, allowing performance optimizations like deduplication and QoS to be applied consistently across different hypervisor types without requiring hypervisor-specific implementations.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS9652265B1Architecture for managing I/O and storage for a virtualization environment with multiple hypervisor types
Publication Date: 2017.05.16 NUTANIX INC
  • US9652265B1 patent drawing
  • US9652265B1 patent drawing
  • US9652265B1 patent drawing

AI summary

A system for managing storage devices, includes a plurality of nodes that implement a virtualization environment, each node of the plurality of nodes comprising a hypervisor, a service virtual machine that sits above the hypervisor, and one or more user virtual machines that sit above the hypervisor; a plurality of storage devices that are accessed by the user virtual machines via the service virtual machines, wherein a first node of the plurality of nodes comprises a first hypervisor, a first service virtual machine and a first set of one or more user virtual machines, wherein a second node of the plurality of nodes comprises a second hypervisor, a second service virtual machine and a second set of one or more user virtual machines, wherein the first hypervisor and the second hypervisor are of different types, and wherein the first virtual machine and the second service virtual machine are of the same type.