Hypervisor Data Chunk Sharing for Virtual Machine Deployment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtualization environments, hypervisors waste resources by independently loading and retrieving common data for multiple virtual machines, as they do not identify and share data chunks derived from a master image, leading to inefficient use of processing power, memory, storage, and network bandwidth.
Innovation Solution
A method where mapping information from a data storage system is used to identify and share data chunks across virtual machines by associating each chunk with a unique identifier, allowing the hypervisor to determine if a copy has already been loaded into memory, and if so, to share or retrieve it only once, using a space-efficient storage scheme like copy-on-write or de-duplication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the hypervisor retrieves and loads each VM image independently into separate memory regions, then each virtual machine can be deployed, but resources are wasted by repeatedly retrieving and loading common data
Solution Approach 1:
The patent merges identical data chunks from multiple VM images into a single copy in memory. By identifying common data across different VM images and consolidating them, the system eliminates redundant loading operations and reduces resource consumption while maintaining VM deployment functionality
Solution Approach 2:
The patent creates a universal data storage approach where a single data chunk can serve multiple VM images simultaneously. The hypervisor loads each unique data chunk once and makes it available to all VM images that require it, enabling one resource to fulfill multiple functions
2Ease of manufacture
If the hypervisor loads all VM images independently without identifying common data, then deployment is straightforward, but storage space is wasted by maintaining multiple copies of identical data
Solution Approach 1:
The patent merges identical data chunks from multiple VM images into a single copy in memory. By identifying common data across different VM images and consolidating them, the system eliminates redundant loading operations and reduces resource consumption while maintaining VM deployment functionality
Solution Approach 2:
The patent uses copying mechanisms to create references to shared data chunks rather than loading complete independent copies. The hypervisor loads each unique data chunk once and creates multiple references to it, reducing memory space while enabling straightforward VM deployment
Data Source
AI summary
Machines, systems and methods for deploying one or more virtual machines on a host computing system, the method comprising: receiving mapping information from a data storage system, wherein the mapping information associates a first data chunk stored in the data storage system with a unique identifier; utilizing the mapping information to determine whether any copies of the first data chunk have already been loaded into a memory of the host computing system in association with deployment of a first virtual machine or a second virtual machine on the host computing system; and in response to determining that no copies of the first data chunk have already been loaded into the memory: retrieving the first data chunk from the data storage system; loading the first data chunk into the memory; and utilizing the first data chunk to deploy the first virtual machine on the host computing system.