Mirage Image Format Decouples Virtual Machine Metadata
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Virtual machine image sprawl occurs due to the need to maintain full stacks of software in each image, leading to storage and management challenges, as each image must be stored and updated independently, despite containing redundant files and metadata.
Innovation Solution
The Mirage Image Format (MIF) decouples the implicit mapping of file names to file content and metadata, creating a manifest that maps file names to unique IDs and a store that maps these IDs to file content, allowing for efficient storage and management of virtual machine images by treating them as structured data, enabling operations without starting the image and reducing data transfer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If virtual machine images are stored as complete files with implicit mappings, then the images can be treated as simple data and are easy to clone and version, but the storage space required increases significantly and management becomes difficult due to redundancy
Solution Approach 1:
The patent segments the virtual machine image into two separate components: a manifest file containing metadata and mappings, and a store containing the actual file content. This segmentation allows the system to store only unique content once in the store while multiple images reference the same content through the manifest, dramatically reducing storage space while maintaining the ability to easily clone and version images by copying references rather than content.
Solution Approach 2:
Instead of copying actual file content when creating image versions or clones, the system copies only the manifest file which contains references to the content store. This allows rapid cloning and versioning operations with minimal storage overhead, as the content itself is not duplicated but merely referenced multiple times through different manifest entries.
2Reliability
If each virtual machine image contains a full stack of software, then the images provide complete isolation and configuration, but the complexity of maintaining and updating each image increases
Solution Approach 1:
By separating the software stack into a standardized base image (stored once in the content store) and image-specific customizations (tracked in the manifest), the system maintains complete isolation and configuration while reducing maintenance complexity. Updates can be applied to the base image once and automatically propagated to all derived images through the reference structure, rather than requiring separate maintenance of each full image stack.
Solution Approach 2:
The content store serves as a universal repository that multiple virtual machine images can reference simultaneously. This multi-functional structure allows a single copy of software to support multiple images, reducing the overall complexity of maintaining software stacks across numerous images while preserving the completeness of each individual image's software environment.
3Loss of information
If complete virtual machine images are transferred and handled for software operations, then all necessary files and metadata are available, but the data transfer time and processing speed decrease
Solution Approach 1:
The patent extracts the essential metadata and mapping information into a separate manifest file that can be transferred independently of the actual content. This allows software operations to be performed by transferring only the lightweight manifest file rather than complete images, dramatically reducing transfer time while maintaining access to all necessary file and metadata information through the manifest's references to the content store.
Solution Approach 2:
The manifest file acts as an intermediary between the user and the content store. It contains all the information needed to identify and access files without requiring the actual file content to be present in memory. This intermediary structure enables fast operations by working with small manifest data while the actual content remains efficiently stored in the content store, reducing both transfer time and processing overhead.
Data Source
AI summary
A method for performing software operations on virtual machine images, includes: receiving a first virtual machine image that includes a mapping from a first file name to first file content; reformatting the first virtual machine image by: creating a unique ID for the first file content; and decoupling the mapping of the first virtual machine image by creating an image manifest that maps the first file name to the unique ID, and creating a store that contains the first file content and that maps the unique ID to the first file content; and performing a software operation on the first virtual machine image using the reformatted first virtual machine image.