Virtual Clock Frequency Control for VM Resource Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual machine systems face challenges in managing resource allocation during tight time-of-day synchronization, leading to potential overloading when multiple tasks are scheduled simultaneously, which can strain system resources and result in failures.
Innovation Solution
Implementing virtualization of real-time clocking by adjusting the frequency of virtual clocks based on specified profiles that define validity periods and allowable time variations, allowing the hypervisor to manage resource distribution efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If tight time-of-day synchronization is maintained across virtual machines, then temporal accuracy is improved, but system resource overload occurs when multiple tasks are scheduled simultaneously
Solution Approach 1:
The patent applies local quality by allowing different virtual machines to have different clock synchronization characteristics based on their specific needs. Each VM can be configured with its own time drift tolerance level, enabling critical VMs to maintain tight synchronization while less critical VMs allow greater drift, thus distributing resource load unevenly across VMs based on their quality requirements
Solution Approach 2:
The patent changes the parameter of clock frequency from a fixed value to a dynamically adjustable parameter. The hypervisor can modify the frequency of virtual clocks within a range, allowing it to slow down or speed up virtual time passage to prevent task scheduling conflicts and resource overload while maintaining acceptable temporal accuracy for each VM
2Productivity
If multiple tasks are scheduled to run at the same time in virtual machines, then productivity is improved, but resource overloading occurs causing system failures
Solution Approach 1:
The patent implements periodic action by introducing controlled time drift and offset variations in virtual clock synchronization. Instead of maintaining perfectly periodic task execution across all VMs, the system allows periodic variations in virtual time passage, which naturally distributes task execution times and prevents simultaneous resource demands while maintaining overall system productivity
Solution Approach 2:
The patent applies dynamics by making the virtual clock frequency adjustable rather than fixed. The hypervisor can dynamically change the frequency of virtual clocks based on current system load conditions, allowing the system to adapt its task scheduling rhythm in real-time to prevent overloading while maintaining high productivity during capacity periods
3Reliability
If virtual clock frequency is adjusted to spread tasks over time, then resource management is improved, but time synchronization accuracy deteriorates
Solution Approach 1:
The patent applies partial action by adjusting virtual clock frequency only when and where necessary to prevent resource overloading. Instead of uniformly degrading time accuracy across all VMs, the hypervisor selectively applies frequency adjustments only to specific VMs experiencing scheduling conflicts, maintaining partial time synchronization accuracy in unaffected VMs
Solution Approach 2:
The patent changes the clock frequency parameter within a controlled range, allowing it to vary from the nominal value but not diverge completely. This parameter change enables resource management flexibility while constraining the degradation of time synchronization accuracy to acceptable levels, balancing resource efficiency with temporal precision
Data Source
AI summary
Examples described herein provide a computer-implemented method that includes receiving a plurality of profiles corresponding to at least one of a plurality of virtual machines. Each of the plurality of profiles defines a validity period and an allowable variation of time. The method further includes determining, based at least in part on the validity period, a subset of the plurality of profiles that are active. The method further includes calculating a uniform distribution for virtual clocks of each of the subset of the plurality of virtual machines and determining a desired offset. The method further includes adjusting a virtual clock frequency of a virtual clock based at least in part on the desired offset and the allowable variation of time. The adjusting causes a difference between a real time-of-day clock and a virtual time-of-day clock of the at least one of the virtual clocks to increase.


