Cloud Instance Recommender for HPC Applications
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to recommend the best cloud setup for High Performance Computing (HPC) applications, as they do not effectively assess cloud readiness and optimize cloud service combinations for HPC workloads.
Innovation Solution
A processor-implemented method and system that assesses cloud readiness of HPC applications by generating a suitability score using application profiling data and cloud instance features, classifying cloud instances, predicting execution time and cost, and recommending optimal cloud instances based on user inputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional recommendation methods are used for general applications, then they can provide basic cloud instance recommendations, but they fail to recommend best cloud setup for HPC applications due to lack of HPC-specific optimization
Solution Approach 1:
The system transforms the recommendation approach by changing parameters from general application metrics to HPC-specific parameters including execution time, cost, suitability score, and performance metrics. This parameter transformation enables accurate recommendations tailored to HPC workload characteristics rather than generic application requirements
Solution Approach 2:
The recommendation system segments cloud instances into different suitability categories (best-suited, average-suited, worst-suited) based on HPC-specific evaluation. This segmentation allows users to identify optimal instances for HPC workloads by dividing the instance pool into performance-based groups rather than treating all instances uniformly
2Manufacturing precision
If users manually assess cloud readiness and optimize cloud service combinations for HPC applications, then they can find optimal configurations, but the process becomes time-consuming and complex
Solution Approach 1:
The system performs preliminary assessment of cloud readiness by pre-calculating suitability scores, execution times, and costs for multiple cloud instances before actual migration. This preliminary action includes generating application profiles, evaluating instance compatibility, and predicting performance metrics in advance, so users make informed decisions without time-consuming manual trials
Solution Approach 2:
The system creates copies of application profiling data and cloud instance specifications to enable virtual evaluation of multiple configurations. By copying and analyzing instance profiles rather than physically migrating and testing each configuration, the system optimizes cloud service combinations without time-consuming actual migrations
3Ease of manufacture
If users try multiple cloud instance options to find the best configuration, then they can identify optimal settings, but users get overwhelmed by different options available to try
Solution Approach 1:
The system provides feedback by ranking cloud instances according to suitability scores and categorizing them into best-suited, average-suited, and worst-suited groups. This feedback mechanism guides users toward optimal instances without requiring them to evaluate all options, reducing overwhelm while maintaining ease of selection
Solution Approach 2:
The system extracts and highlights only the most relevant cloud instances for HPC workloads by filtering out unsuitable options. By taking out and presenting only the best-suited instances based on HPC-specific criteria, the system reduces the number of options users need to consider while maintaining ease of selection
4Productivity
If HPC applications are executed on dedicated on-premise clusters, then performance requirements are met, but cost and management complexity increase
Solution Approach 1:
The recommendation system enables cloud instances to serve multiple HPC workload types by identifying instances with universal suitability for different HPC applications. By finding instances that can handle various HPC workloads effectively, the system replaces dedicated on-premise clusters with flexible cloud resources that maintain productivity while reducing management complexity
Data Source
AI summary
The present disclosure discloses a method and system for generating recommendations for cloud instances for high performance computing (HPC) applications. The present disclosure provides an intelligent Cloud instance Recommender framework comprising a suitability matcher, a performance analyzer, and a decision making enabler. The method of the present disclosure ensures that the HPC application is assessed for its suitability for the cloud since there is no need of recommending cloud services if the HPC application cannot be migrated to the cloud. This assessment is performed using a machine learning (ML) predictor engine which is trained upon some parameters of the HPC application. The ML predictor engine predicts execution time of the HPC application on cloud instances, and then a cost of execution is estimated by a mathematical model based on the predicted execution time. Also, a weightage to user's input is provided using a recommender engine to generate final recommendations.