ML-Based Cloud Resource Customization for User Satisfaction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud resource management systems face challenges in optimizing resource allocation, leading to user dissatisfaction due to misbehaved applications, resource wastage, and improper handling of prospective customers, as they lack dynamic and personalized approaches to predict user behavior and resource needs.
Innovation Solution
A computer-implemented method using machine learning to evaluate current users based on attributes, predict user value scores, cluster users, and customize resource allocation for applications by assessing the likelihood of misbehavior and future paying potential, ensuring balanced resource utilization and user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If cloud resources are allocated uniformly to all users, then resource management is simple, but user satisfaction decreases due to inability to address individual user needs and misbehaved applications
Solution Approach 1:
The patent segments users into different groups (current users, prospective users, high-value users) and applications into different categories (misbehaved, normal) based on machine learning predictions. This segmentation enables differentiated resource allocation strategies for different user segments, improving user satisfaction while maintaining manageable complexity through automated classification.
Solution Approach 2:
The system dynamically changes resource allocation parameters based on predicted user value scores and application behavior likelihoods. Machine learning models continuously update predictions about user conversion probability and misbehavior risk, allowing the system to adjust resource allocation parameters in real-time to optimize both user satisfaction and resource efficiency.
2Productivity
If resources are allocated based on predicted user value, then resource utilization efficiency improves, but measurement and prediction accuracy requirements increase
Solution Approach 1:
The system implements feedback loops where machine learning models continuously learn from actual user behavior outcomes to improve prediction accuracy. Resource allocation decisions based on predictions generate new data that feeds back into the models, progressively enhancing measurement precision for user value assessment and enabling more efficient resource utilization over time.
Solution Approach 2:
The system performs preliminary resource allocation based on predicted user value before actual user behavior is fully observed. By making predictions about prospective users' conversion likelihood and current users' continued value, the system can pre-allocate resources optimally, improving overall resource utilization efficiency while the predictions continue to refine with new data.
3Loss of energy
If resources are restricted for prospective users, then resource waste from non-converting users decreases, but conversion opportunities may be lost
Solution Approach 1:
The patent applies local quality by treating different prospective user segments differently based on their predicted conversion probability. High-value prospective users with high conversion likelihood receive adequate resource allocation to maintain conversion opportunities, while low-probability users receive restricted resources to minimize waste. This differentiated approach optimizes both resource efficiency and conversion rates.
Solution Approach 2:
The system dynamically changes resource allocation parameters for prospective users based on predicted conversion probability. Machine learning models continuously update these predictions, allowing the system to adjust resource parameters in real-time - restricting resources for low-probability users to reduce waste while maintaining or increasing resources for high-probability users to preserve conversion opportunities.
4Measurement precision
If machine learning predictions are used for resource allocation, then resource allocation accuracy improves, but system complexity and computational requirements increase
Solution Approach 1:
The machine learning system performs self-service by automatically evaluating users, predicting their value, and making resource allocation decisions without manual intervention. The models continuously train on new data and autonomously adjust allocation parameters, improving accuracy while managing complexity through automation rather than manual processes.
Solution Approach 2:
The patent employs universal machine learning models that serve multiple functions: predicting user conversion probability, assessing continued user value, identifying misbehaved applications, and guiding resource allocation decisions. This multi-functionality improves resource allocation accuracy across different user types and scenarios while avoiding the need for separate complex systems for each function.
Data Source
AI summary
Customizing computing resource allocation based on machine learning is provided. A plurality of current users of a service are evaluated based on attributes of each current user. A user value score is generated for each current user based on the attributes. The plurality of current users is clustered into a plurality of user groups based on the user value score. A likelihood of each prospective user becoming a paying customer is predicted based on collected behavior data corresponding to each prospective user. A likelihood of an application corresponding to the service and associated with a particular user will misbehave is predicted based on resource usage and features of the application. Resource allocation to the application corresponding to the service and associated with the particular user is customized based the likelihood that the application will misbehave and whether the particular user is a current user or a prospective user.