Cloud Service Instance Allocation via Resource Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deploying applications or updates in cloud-computing environments can lead to performance issues due to resource insufficiencies, which are often undetected by simple function/load tests, especially in micro-service based deployments, resulting in resource waste and lack of risk alert.
Innovation Solution
A method that determines service instance counts and host instance counts based on resource capacity models and historical consumption data, allocating services to host devices based on resource correlation, allowing for optimal resource allocation and migration to ensure efficient resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple services are deployed in one single host device to improve deployment speed, then productivity is improved, but resource insufficiency problems occur that are undetected by simple tests
Solution Approach 1:
The system performs preliminary resource correlation analysis and simulation testing before actual deployment. By analyzing historical resource consumption data and calculating service similarity in advance, the system predicts potential resource conflicts before they occur, allowing preventive measures to be taken while maintaining fast deployment speeds.
Solution Approach 2:
The system implements a feedback mechanism where resource consumption data from deployed services is continuously monitored and fed back to the deployment system. This feedback loop enables the system to detect resource insufficiency issues in production environments and adjust future deployment decisions accordingly, improving reliability without sacrificing deployment speed.
2Manufacturing precision
If resource capacity models and historical consumption data are analyzed to determine optimal service instance counts, then manufacturing precision is improved, but loss of time increases due to additional analysis requirements
Solution Approach 1:
The system pre-calculates service similarity metrics and resource correlation coefficients during off-peak periods or as part of continuous background processing. By having this analysis ready in advance, the system can quickly determine optimal service instance counts during deployment without adding significant time overhead, thus maintaining high precision while minimizing time loss.
Solution Approach 2:
The system dynamically adjusts the depth of analysis based on deployment context. For routine deployments, it uses pre-computed similarity metrics and standard resource models. For critical or unusual deployments, it performs more comprehensive analysis. This parameter-based approach allows the system to balance precision and time loss flexibly according to actual needs.
3Loss of energy
If services are allocated based on service similarity and resource correlation, then resource waste is reduced, but device complexity increases due to sophisticated allocation algorithms
Solution Approach 1:
The system uses service similarity analysis to identify groups of services with comparable resource consumption patterns. Instead of treating each service individually, it creates representative models or 'copies' of service groups and allocates resources based on these models. This approach significantly reduces computational complexity while still achieving efficient resource utilization and minimizing waste.
Solution Approach 2:
The system develops a universal resource correlation model that can be applied across different service types and deployment scenarios. This multi-functional model handles various service characteristics (CPU-intensive, memory-intensive, I/O-bound) through a unified framework, reducing the need for service-specific complex algorithms and simplifying the overall allocation system.
Data Source
AI summary
A method, computer program product, and computer system for determining, by a computing device, a service instance count for each of a plurality of services to be executed on a plurality of host devices. A similarity between the plurality of services may be determined. A host instance count for the plurality of host devices may be determined based upon, at least in part, the similarity between the plurality of services. Each service instance may be allocated to a respective host device for execution based upon, at least in part, the similarity between the plurality of services, the service instance count, and the host service count.