Hypervisor Credit Scheduler for Soft Real-Time Task Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hypervisors struggle to effectively manage and balance the execution of time-sensitive and non-time-sensitive tasks in data-processing systems, leading to performance degradation, especially in voice and media-related applications, due to inadequate resource allocation and prioritization.
Innovation Solution
An enhanced hypervisor that dynamically adjusts scheduling parameters based on performance metrics such as total-time, timeslice, total-latency, credit-latency, and average-latency to optimize task execution, using a credit scheduler that monitors and tunes queue priorities and resource allocation in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a hypervisor schedules tasks without dynamic adjustment, then the system structure is simple, but time-sensitive tasks experience performance degradation due to inadequate resource allocation
Solution Approach 1:
The patent implements dynamic scheduling by continuously monitoring performance metrics (total-time, timeslice, total-latency, credit-latency, average-latency) and adjusting credit allocations in real-time. The hypervisor transitions from static scheduling to dynamic adjustment, where scheduling parameters are modified based on actual task performance data to optimize resource distribution for time-sensitive tasks.
Solution Approach 2:
The patent establishes a feedback mechanism where performance metrics are collected, analyzed, and used to adjust credit allocations. The system monitors execution time, latency, and timeslice data, compares it against thresholds, and automatically modifies scheduling parameters accordingly. This closed-loop feedback enables continuous optimization of resource allocation for time-sensitive tasks.
2Loss of time
If resources are allocated statically to domains, then the scheduling mechanism is simple, but latency increases for time-sensitive tasks
Solution Approach 1:
The system transitions from static resource allocation to dynamic credit allocation that responds to real-time performance conditions. Credits are adjusted based on monitored metrics including total-latency and average-latency, enabling the system to reduce latency for time-sensitive tasks by allocating more resources when needed and fewer when available.
Solution Approach 2:
The patent modifies scheduling parameters (credit allocations, weightings, priorities) based on performance thresholds. When metrics such as total-time or latency exceed thresholds, the system changes allocation parameters to favor time-sensitive tasks. This parameter adjustment mechanism directly reduces latency while managing system complexity through automated decision-making.
3Productivity
If the hypervisor monitors multiple performance metrics in real-time, then task execution is optimized, but processing overhead increases
Solution Approach 1:
The system performs self-monitoring and self-adjustment by having the hypervisor collect its own performance metrics and automatically generate scheduling decisions. This self-service capability eliminates the need for external monitoring systems or manual intervention, reducing overall processing overhead while maintaining optimization benefits.
Solution Approach 2:
The feedback mechanism processes performance metrics efficiently by using threshold-based decision logic rather than complex continuous optimization. The system compares monitored values against predefined thresholds and adjusts credits accordingly, a straightforward approach that minimizes processing overhead while achieving meaningful productivity improvements.
4Reliability
If credit allocations are adjusted frequently, then task performance is optimized, but system stability decreases
Solution Approach 1:
The system uses feedback from performance metrics to adjust credit allocations, but implements stability through threshold-based control. Changes are triggered only when metrics cross predefined thresholds, preventing constant adjustment. This feedback-with-thresholds approach maintains performance optimization while avoiding excessive volatility in scheduling parameters.
Solution Approach 2:
The patent implements periodic monitoring and adjustment cycles rather than continuous modification. The system evaluates performance metrics at regular intervals and adjusts credits based on whether thresholds are crossed during each cycle. This periodic action pattern stabilizes the system by introducing rhythm and predictability to the adjustment process while maintaining optimization benefits.
Data Source
AI summary
Methods to dynamically improve soft real-time task performance in virtualized computing environments under the management of an enhanced hypervisor comprising a credit scheduler. The enhanced hypervisor analyzes the on-going performance of the domains of interest and of the virtualized data-processing system. Based on the performance metrics disclosed herein, some of the governing parameters of the credit scheduler are adjusted. Adjustments are typically performed cyclically, wherein the performance metrics of an execution cycle are analyzed and adjustments may be applied in a later execution cycle. In alternative embodiments, some of the analysis and tuning functions are in a separate application that resides outside the hypervisor. The performance metrics disclosed herein include: a “total-time” metric; a “timeslice” metric; a number of “latency” metrics; and a “count” metric. In contrast to prior art, the present invention enables on-going monitoring of a virtualized data-processing system accompanied by dynamic adjustments based on objective metrics.