Virtual Machine Template Segmentation for Storage Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Provisioning hundreds of thousands of virtual machines in a distributed virtual infrastructure data center is a challenging task due to scalability, reliability, performance, and management issues, particularly in terms of data storage and deduplication.
Innovation Solution
The implementation of a content-centric networking approach, where virtual machine templates are segmented into addressable content portions using a content-centric network communications protocol, allowing for efficient deployment and management of virtual machines by addressing content rather than location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If virtual machine templates are deployed using traditional location-based networking, then deployment speed is maintained, but storage efficiency deteriorates due to inability to deduplicate similar VM images
Solution Approach 1:
The virtual machine template is divided into multiple segments, each representing a content portion that can be independently addressed and stored. This segmentation enables the system to identify and deduplicate identical segments across different VM templates while maintaining the ability to quickly assemble complete VM images when needed, thus resolving the contradiction between storage efficiency and deployment speed.
2Reliability
If content is addressed by location in traditional networks, then network communication is simple, but data management efficiency deteriorates due to inability to locate and share identical content across different locations
Solution Approach 1:
The patent replaces the traditional location-based addressing mechanism (mechanical system) with a content-based addressing system. Instead of locating content by its physical position in the network, the system uses content identifiers to address and retrieve data portions. This substitution enables efficient data management and deduplication while the underlying network infrastructure maintains simple communication protocols.
3Productivity
If multiple virtual machines are provisioned simultaneously using traditional methods, then provisioning throughput increases, but network workload and latency increase due to redundant data transmission
Solution Approach 1:
The patent merges identical content portions from multiple virtual machine templates into a single stored copy. When multiple VMs need to be provisioned simultaneously, the system retrieves the shared content segments once and makes them available to all requesting VMs, thereby maintaining high provisioning throughput while significantly reducing network workload and avoiding redundant data transmission.
4Reliability
If virtual machine templates are stored as complete images, then retrieval is straightforward, but storage efficiency deteriorates due to duplication of identical content across multiple templates
Solution Approach 1:
The patent segments virtual machine templates into smaller content portions that can be independently stored and managed. This segmentation allows the system to store only unique segments once and reference them across multiple templates, improving storage efficiency while maintaining retrieval simplicity through an indexing system that maps template requirements to stored segments.
Solution Approach 2:
The patent introduces an intermediary indexing system that sits between the stored content segments and the template retrieval process. This intermediary maintains mappings between template requirements and stored content portions, enabling straightforward template assembly without requiring complete image storage, thus resolving the contradiction between storage efficiency and retrieval simplicity.
Data Source
AI summary
Techniques are disclosed for provisioning virtual machines via content-centric networking approach in a computing system implemented with a distributed virtual infrastructure. In one example, a method comprises the following steps. A virtual machine template is obtained. The virtual machine template defines one or more segments which represent one or more content portions. The one or more content portions are addressable in accordance with a content-centric network communications protocol. At least one virtual machine is deployed from the virtual machine template. The generating and deploying steps are performed on one or more physical processing elements associated with a computing system implemented via a distributed virtual infrastructure.


