Cloud Workload CPU and Memory Limits with Gradual Percentage Changes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In cloud computing environments, altering resource limits for workloads can be risky and potentially disrupt the performance and stability of other workloads, especially when reducing resources abruptly.
Innovation Solution
Gradually alter CPU and memory limits by a consistent percentage over time, rather than in integer units, to minimize system instability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If resource limits are altered abruptly in integer units, then the change is simple to implement, but system stability and performance are compromised
Solution Approach 1:
The patent segments a single large resource limit change into multiple smaller incremental changes. Instead of changing resource limits by fixed integer units (e.g., reducing CPU from 16 to 1 in one step), the system divides the total change into N smaller steps, each applying a fraction of the total change. This segmentation reduces the impact of each individual change while achieving the same ultimate goal, thereby maintaining system stability during the transition.
Solution Approach 2:
The patent implements dynamic adjustment of resource limits by continuously monitoring system performance metrics during the transition period. The system adapts the rate and magnitude of resource limit changes based on real-time feedback from the workload performance, rather than applying static predetermined changes. This dynamic approach allows the system to slow down or pause adjustments if performance degradation is detected, ensuring reliability while still achieving the desired resource allocation change.
2Productivity
If resource limits are reduced significantly to improve resource allocation efficiency, then more resources become available for other workloads, but performance and stability of existing workloads are impacted
Solution Approach 1:
The patent applies preliminary action by implementing resource limit changes gradually over a specified transition period before the full reduction takes effect. During this transition period, the system applies incremental reductions (e.g., reducing by 10% each day over 10 days) rather than implementing the full reduction immediately. This allows the workload to adapt to the changing resource constraints progressively, maintaining performance while eventually achieving the improved resource allocation efficiency.
Solution Approach 2:
The patent incorporates feedback mechanisms that continuously monitor workload performance metrics during the resource limit adjustment process. The system measures performance indicators (such as response time, throughput, or error rates) and uses this feedback to adjust the rate of resource limit reduction. If performance degradation exceeds thresholds, the system can slow down or pause the reduction, ensuring that productivity improvements do not come at the cost of workload reliability.
3Loss of time
If resource limits are changed in large increments, then the transition to target allocation is faster, but the risk of system instability increases
Solution Approach 1:
The patent segments the transition process into multiple small incremental steps rather than applying large changes in few steps. By dividing the total resource limit change into N smaller increments and applying them over an extended period, the system reduces the magnitude of each individual change event. This segmentation minimizes the risk of system instability while still achieving the transition, as each small change is less likely to trigger performance degradation or system failures.
Solution Approach 2:
The patent implements periodic action by applying resource limit changes at regular intervals over a specified transition period. Instead of making one large change or irregular small changes, the system applies consistent incremental adjustments at predetermined time intervals (e.g., daily reductions during a 10-day transition). This periodic approach provides predictability and allows the system to maintain stability through rhythmic, controlled adjustments rather than abrupt or erratic changes.
Data Source
AI summary
A computer implemented method, system and product for gradually altering compute resources allocated to workloads in a cloud computing environment is described. A fixed percentage per increment is determined by which to alter a current allocation of at least one of a set of compute resources allocated to at least one of a set of workloads in a cloud computing environment during a selected period of time to incrementally switch the current allocation of the at least one of the set of compute resources allocated to the at least one of the set of workloads in the cloud computing environment from an original allocation to a target allocation. The current allocation is switched incrementally by the fixed percentage per increment the current allocation from the original allocation to the target allocation during the selected period of time.