Cloud Service Benchmarking for Cross-CSP Performance-Cost Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud service comparison and evaluation systems are limited to intra-CSP comparisons, lacking the ability to accurately quantify performance and cost differences between heterogeneous cloud service providers, making it difficult for users to optimize their cloud service usage based on quantitative numerical values and discount policies.
Innovation Solution
A system and method that standardizes performance and cost evaluation across different cloud service providers by measuring actual performance and discount costs, enabling comparison and optimization of cloud services using a cloud service management server with benchmarking, cost calculation, and service optimization modules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If cloud service providers provide detailed performance information (number of cores, threads), then users can see basic specifications, but accurate performance comparison between heterogeneous CSPs remains impossible
Solution Approach 1:
The patent introduces a cloud service management server as an intermediary that acts as a mediator between users and multiple cloud service providers. This server benchmarks actual performance of cloud services from different CSPs and provides standardized conversion numerical values, enabling accurate cross-CSP performance comparison without requiring users to directly analyze complex provider-specific performance data.
Solution Approach 2:
The patent transforms diverse performance parameters (cores, threads, CPU architecture-specific metrics) from different cloud service providers into a unified set of standardized conversion numerical values through benchmarking. This parameter transformation enables apples-to-apples comparison across heterogeneous cloud services by converting them into a common measurement framework.
2Loss of energy
If cloud service providers offer various discount policies (spot instance, savings plan, reserved instance), then users can reduce costs, but comparing and selecting optimal discounts across different CSPs becomes difficult
Solution Approach 1:
The cloud service management server serves as an intermediary that collects and processes discount policy information from multiple CSPs. It calculates actual costs after applying discounts and presents standardized cost comparisons, making it easier for users to evaluate and select optimal discount policies across different providers without having to manually analyze each CSP's specific pricing structures.
Solution Approach 2:
The patent creates a universal evaluation framework that handles multiple types of discount policies (spot instances, savings plans, reserved instances) from different cloud service providers using the same standardized comparison methodology. This universal approach allows users to compare heterogeneous discount offerings across CSPs using a single consistent evaluation process.
3Loss of energy
If users want to select cheap cloud services with desired performance, then cost efficiency improves, but accurate performance-to-cost comparison across heterogeneous CSPs remains difficult
Solution Approach 1:
The patent transforms performance parameters into standardized conversion numerical values through systematic benchmarking. By converting all cloud service performance metrics into this unified standard, the system enables precise performance-to-cost ratio calculations across heterogeneous CSPs, allowing users to accurately compare cost efficiency regardless of provider or service type.
Solution Approach 2:
The patent replaces manual, subjective performance estimation with an automated benchmarking system that objectively measures and converts performance metrics. This substitution of mechanical measurement processes enables consistent, repeatable, and accurate performance-to-cost comparisons that eliminate human bias and estimation errors.
4Adaptability or versatility
If cloud service providers use different CPU architectures and specifications, then service diversity and customization improve, but performance quantification and comparison between services becomes inaccurate
Solution Approach 1:
The patent applies parameter transformation by converting architecture-specific performance metrics into standardized conversion numerical values. Whether measuring AMD EPYC, Intel Xeon, or other CPU architectures through different performance characteristics, the system transforms all these into a unified measurement scale that maintains accuracy while enabling cross-architecture comparison.
Solution Approach 2:
The patent creates virtual benchmarks that copy and standardize performance measurements across different hardware architectures. By establishing standardized benchmarking procedures that can be replicated across various CPU types and cloud service configurations, the system maintains the diversity of service offerings while achieving consistent, comparable performance quantification.
Data Source
AI summary
A system for managing and evaluating a cloud service includes a user terminal through which a user is provided with at least one cloud service, at least one cloud service provider (CSP) server configured to provide the user with the at least one cloud service through the user terminal, and a cloud service management server configured to communicate with the at least one CSP server and to compare at least one existing cloud service provided from the at least one CSP server to the user terminal and at least one candidate cloud service capable of being provided to the user terminal.