Virtualization Scaling Management via Dynamic Method Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current virtualization infrastructure faces challenges with load balancing between virtual machines or containers, slow image copying, long configuration loading times, and difficulty in achieving fine-granularity scaling, particularly in scale-out operations which can lead to inefficient resource allocation and increased scaling times.
Innovation Solution
A virtualization infrastructure that includes a resource monitoring unit, scaling execution determination unit, resource dispensing control unit, and scaling method determination unit to dynamically select between scale-out, scale-up, and combined scale-out/scale-up methods based on available resources, allowing for high-speed and fine-granularity scaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If scale-out is used to increase the number of virtual machines or containers, then processing capacity is optimized, but image copying time increases and activation time increases
Solution Approach 1:
The patent applies preliminary action by pre-copying images to edge cache memory before scaling operations are needed. The image management unit copies images from storage to edge cache in advance, so when scale-out is required, the images are already available locally, eliminating the delay of copying images during the scaling operation itself.
Solution Approach 2:
The patent uses copying by creating virtual machine or container instances from stored images. The image management unit copies images to edge cache, and the virtual machine management unit creates multiple instances from these pre-cached images, enabling rapid scale-out without repeated image copying from remote storage.
2Productivity
If scale-out is used to increase the number of virtual machines or containers, then processing capacity is optimized, but configuration loading time increases
Solution Approach 1:
The patent applies preliminary action by pre-loading configuration files into the edge cache along with images. The image management unit retrieves and caches both images and their associated configuration files before scaling operations, so when new virtual machines or containers are created, the configuration is already available locally and can be loaded immediately without waiting for remote retrieval.
3Manufacturing precision
If fine-granularity scaling is implemented, then resource allocation precision is improved, but the number of scaling operations increases causing longer total scaling time
Solution Approach 1:
The patent merges scale-up and scale-out operations into a unified scaling process. When resources are needed, the system can simultaneously allocate additional resources to existing virtual machines or containers (scale-up) and create new instances (scale-out) from pre-cached images and configurations, reducing the total number of separate scaling operations and their cumulative time.
Solution Approach 2:
The system implements self-service by automatically selecting the appropriate scaling method (scale-up or scale-out) and executing the scaling operation without manual intervention. The scaling management unit monitors resource usage and automatically triggers scaling when thresholds are met, reducing delays associated with manual scaling decisions and executions.
4Adaptability or versatility
If hot plug is used to flexibly add or remove CPU, memory, and disk, then resource flexibility is improved, but the virtual machine must be operating on a single compute with excessive resource
Solution Approach 1:
The patent applies dynamics by enabling resource allocation to change dynamically based on actual demand. The system monitors resource usage and automatically scales resources up or down, and can migrate virtual machines between computes, ensuring resources are allocated precisely when and where needed rather than being statically over-provisioned on a single compute.
Solution Approach 2:
The patent segments resources across multiple computes rather than concentrating them on a single compute. By distributing virtual machines and resources across multiple compute nodes, the system achieves flexibility through load balancing and migration capabilities, allowing resources to be dynamically reallocated without requiring excessive capacity on any single compute.
Data Source
AI summary
[Problem] To realize high-speed and fine-granularity scaling in a virtualization infrastructure.[Solution] A virtualization infrastructure 1 includes: a monitoring unit that monitors loads on a plurality of computes and loads on virtual machines that are execution environments operating on the plurality of computes; a scaling execution determination unit that determines execution of scaling and calculates a required resource amount; a resource dispensing control unit that manages and allocates a resource; a scaling method determination unit that, upon the determination of execution of scaling by the scaling execution determination unit, checks availability of the resource in each of the plurality of computes and selects any scaling method from the group consisting of scale-out, scale-up, and combined use of scale-out and scale-up; and a scaling indication unit that indicates the selected scaling method selected by the scaling method determination unit.