Cloud Resource Allocation via User Predictability Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud computing resource schedulers primarily rely on current resource utilization metrics like CPU, RAM, and disk capacity, failing to redistribute resources during operation and neglecting the type of applications running on the same cluster, leading to unsatisfactory performance.
Innovation Solution
A method and apparatus that determine a prediction accuracy score for each user based on their previous usage patterns to allocate cloud computing resources, creating groups of users with different predictabilities and applying system policies for optimal resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If resource schedulers rely on current resource utilization metrics (CPU, RAM, disk capacity), then resource allocation decisions can be made quickly based on available data, but the system cannot redistribute resources during operation and does not account for application types, leading to suboptimal performance
Solution Approach 1:
The system performs preliminary analysis of historical usage data to determine predictability scores for each user before allocating resources. By pre-calculating these scores based on past behavior patterns, the system can make more informed allocation decisions that consider both current demand and historical reliability, enabling resource redistribution during operation while maintaining allocation efficiency
2Device complexity
If resource schedulers use current utilization metrics only, then the system structure remains simple, but the system fails to redistribute resources during operation and cannot adapt to different application types
Solution Approach 1:
The system pre-calculates predictability scores for users based on historical usage patterns before resource allocation decisions are made. This preliminary characterization of user behavior enables the scheduler to adapt to different application types and redistribute resources during operation without fundamentally redesigning the scheduler architecture, thus maintaining structural simplicity while gaining adaptability
Solution Approach 2:
The system dynamically adjusts resource allocation based on real-time requests while incorporating static predictability scores derived from historical data. This allows the scheduler to adapt to changing workload patterns and application types during operation, enabling resource redistribution without requiring complete redesign of the allocation mechanism
3Ease of operation
If all users are treated equally in resource allocation, then the system is simple to implement, but unpredictable users can degrade the performance of predictable users
Solution Approach 1:
The system applies different allocation strategies to different users based on their predictability scores. By characterizing each user's historical behavior patterns, the system can tailor resource allocation decisions to individual user characteristics, allowing predictable users to receive allocations optimized for their consistent patterns while unpredictable users are managed separately, thus protecting performance consistency without requiring complete policy redesign
Data Source
AI summary
A method of allocating cloud computing resources in a cloud computing environment to a user is disclosed. The method comprises determining a prediction accuracy score for each user indicative of their user predictability, and allocating cloud computing resources to each user dependent on their user predictability.