Computational Service Pricing for Runtime-Cost Tradeoffs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current software licensing models for modeling and simulation (M&S) software are inflexible and lack the ability to optimize performance based on user needs or budget, leading to inefficient use of cloud computing resources and lack of domain expertise required for effective performance tuning.
Innovation Solution
A computer-implemented system and method that provides a range of prices for predicted performance wall-clock computing times, allowing users to optimize resource allocation based on their performance requirements and budget through a resource estimation model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional licensing models (perpetual or time-based) are used for M&S software, then software access is provided, but flexibility to optimize performance based on user needs or budget is lost
Solution Approach 1:
The patent implements dynamic licensing that allows users to adjust their software usage rights based on changing performance needs and budget constraints. The system continuously monitors resource utilization and automatically adjusts license allocations, transforming the static traditional licensing model into a dynamic one that adapts to user requirements in real-time.
Solution Approach 2:
The system enables users to modify key parameters such as computational resources, time periods, and performance levels of their software licenses. By allowing parameter changes without requiring new license purchases, the system provides flexibility while maintaining ease of operation through standardized adjustment mechanisms.
2Productivity
If cloud computing resources are scaled up to improve simulation throughput, then computational capacity increases, but cost control becomes difficult
Solution Approach 1:
The patent implements a feedback mechanism that continuously monitors cloud resource utilization, simulation throughput, and associated costs. This feedback loop provides users with real-time information about resource consumption patterns, enabling them to optimize their computational resource allocation to achieve desired throughput while maintaining cost control through data-driven decision-making.
Solution Approach 2:
The system allows users to allocate computational resources on-demand, enabling partial utilization of cloud resources rather than requiring full-scale deployments. Users can start with minimal resources and incrementally increase capacity only when and where needed, avoiding excessive resource allocation and associated costs.
3Adaptability or versatility
If usage-based licensing model is implemented, then flexibility for surge-style M&S needs is improved, but ability to determine cost a priori is lost
Solution Approach 1:
The patent implements preliminary cost estimation functionality that calculates projected costs before users commit to usage-based licensing arrangements. The system analyzes historical usage patterns, current resource requirements, and pricing structures to provide accurate cost predictions, enabling users to plan their budgets in advance while retaining the flexibility of usage-based licensing for surge needs.
4Speed
If more computational resources are allocated, then simulation performance improves, but understanding of diminishing returns requires domain expertise
Solution Approach 1:
The patent implements automated performance optimization that uses machine learning algorithms to analyze the relationship between computational resource allocation and simulation performance. The system automatically identifies the optimal resource allocation point where diminishing returns set in, eliminating the need for users to possess domain expertise in performance tuning while maximizing simulation speed and efficiency.
Data Source
AI summary
A computer-implemented method for providing a range of prices for a range of predicted performance wall-clock computing times for executing a desired software operation. A user uploads data of a desired software operation to a service provider. The service provider analyzes the data using a resource estimation model, determines resource options, and converts the resource options into a range of prices and corresponding predicted performance wall-clock times for execution of the desired software operation. The user selects a particular price and corresponding predicted performance wall-clock time and requests execution of the selected software operation using the resource option associated with the selected price and corresponding wall-clock time. An opportunity is provided to the user to optimize price based on either a need for more timely solutions or a need to minimize costs.