Virtual Machine Allocation via Differential Disk Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale thin client systems with thousands to tens of thousands of virtual machines face difficulties in producing standby systems and executing maintenance efficiently due to the complexity of reverting processing across shared virtual machines.
Innovation Solution
A thin client system architecture that includes a connection management server with a ranking information storage unit, session management unit, and maintenance processing management unit, which prioritizes virtual machines for reverting based on differential disk capacity, allowing efficient allocation and maintenance of virtual machines by ranking them for reverting processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If virtual machine duplexing is used for maintenance, then system availability is maintained, but it becomes difficult to produce standby systems and execute maintenance efficiently in large-scale systems with thousands to tens of thousands of virtual machines
Solution Approach 1:
The patent segments the large-scale virtual machine maintenance problem into manageable units by creating a queue-based system that processes virtual machines individually or in controlled batches. The connection management server divides the maintenance task into discrete steps: identifying candidates, queuing them, and executing maintenance one at a time, transforming an overwhelming complex problem into a series of simple, manageable operations.
Solution Approach 2:
The patent introduces a connection management server as an intermediary between the virtual machine management server and the maintenance system. This intermediary component queues maintenance requests and coordinates the maintenance execution, acting as a buffer that simplifies the interaction between multiple system components and enables efficient maintenance without requiring complex direct coordination between all system elements.
2Ease of operation
If all virtual machines are reverted at a fixed maintenance time, then maintenance is simplified, but it causes system-wide impact and reduced availability
Solution Approach 1:
The patent implements preliminary action by pre-identifying and queuing virtual machine candidates for maintenance before the actual maintenance execution. The connection management server maintains a queue of virtual machines that are identified as suitable for maintenance, allowing the system to prepare maintenance candidates in advance without immediately executing maintenance on all machines at once, thus maintaining availability while simplifying maintenance operations.
Solution Approach 2:
The patent implements periodic action by executing maintenance on virtual machines in a staged, periodic manner rather than all at once. The connection management server processes the maintenance queue periodically, selecting and maintaining virtual machines in sequence or in controlled batches over time, which maintains system availability while still achieving comprehensive maintenance coverage.
3Quantity of substance
If virtual machines with small differential disk capacity are prioritized for reverting, then disk space is preserved, but it increases the total reverting time
Solution Approach 1:
The patent applies partial action by selectively maintaining only certain virtual machines based on predefined criteria rather than maintaining all virtual machines uniformly. The connection management server identifies specific virtual machine candidates for maintenance based on factors such as usage patterns, differential disk capacity, and priority levels, performing maintenance partially on a subset of machines rather than comprehensively on all machines, thus optimizing the balance between disk space preservation and reverting time.
Data Source
Figure 1~2
Figure 3~5
Figure 6~8
AI summary
A thin client system comprises a virtual machine server (50), a thin client terminal (40), a connection management server (20), and a virtual machine management server which executes reverting processing of the virtual machine, wherein the connection management server includes a session management unit (11) which determines the virtual machine to be allocated to the thin client terminal based on a to-be-reverted list stored in which virtual machines to be reverted are arranged in descending order of differential disk capacities, and a maintenance processing management unit which requests the virtual machine management server to execute reverting processing, wherein the session management unit when a virtual machine not included in the to-be-reverted list exists, allocates the virtual machine in question to the thin client terminal, and when all the virtual machines are included in the to-be-reverted list, allocates the virtual machine whose differential disk capacity is the smallest to the thin client terminal.