Burstable Instance Performance Reservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud-based service provider networks face challenges in efficiently utilizing computing resources, particularly with burstable instances that may experience CPU throttling due to competing demands, leading to underutilization or over-provisioning of resources.
Innovation Solution
The implementation of a resource service that allows users to reserve increased performance levels for burstable instances for a limited time period, ensuring that computing resources are allocated efficiently and effectively meet workload demands without throttling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If burstable instances share computing resources with multiple other instances to increase utilization, then resource utilization improves, but CPU throttling occurs during peak demand
Solution Approach 1:
The system performs preliminary actions by pre-allocating baseline computing resources to each burstable instance before peak demand occurs. This baseline allocation ensures that when multiple instances compete for shared resources, each instance has a guaranteed minimum level of performance, preventing CPU throttling while still allowing resources to be shared and utilized efficiently during normal operation.
2Reliability
If dedicated computing resources are allocated to each instance to prevent throttling, then performance reliability improves, but resource utilization decreases
Solution Approach 1:
The computing resources are segmented into two distinct parts: baseline resources and burst resources. Baseline resources are dedicated to each burstable instance to guarantee minimum performance and prevent throttling. Burst resources remain shared and available for temporary allocation when instances need additional capacity. This segmentation allows the system to maintain performance reliability through dedicated baseline resources while achieving high utilization through shared burst resources that are allocated on-demand.
3Reliability
If computing resources are over-provisioned to meet peak demand, then performance during peaks improves, but idle resources increase during low demand
Solution Approach 1:
The resource allocation system is made dynamic by allowing burstable instances to flexibly adjust their resource consumption based on actual demand. Instances start with a guaranteed baseline allocation that ensures minimum performance. When peak demand occurs, instances can dynamically burst to access additional shared resources temporarily. When demand decreases, instances return to baseline levels, freeing up resources for other uses. This dynamic approach eliminates the need for static over-provisioning, ensuring performance during peaks while minimizing idle resource waste during low-demand periods.
Data Source
AI summary
This disclosure describes techniques for increasing the baseline performance of a burstable instance to an increased performance level for a limited time period. For example, a user may schedule a time period which the burstable instance has access to 100% of a CPU, instead of competing with other burstable instances even during periods of bursting as in prior techniques. In some cases, the user uses credits to reserve the time periods at which the increased performance level is requested. After receiving the reservation, a resource system selects computing resources to host the burstable instance such that the burstable instance, can operate at the requested increased performance level. After the time period has ended, the burstable instance may return to the baseline performance level.