Reliability-Based Resource Allocation in Cloud Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current resource allocation methods in software-defined infrastructures, such as cloud computing, rely on simplistic threshold-based approaches that fail to consider reliability and cost models, leading to inefficient provisioning and decommissioning of resources.
Innovation Solution
Implementing a resource allocation method based on reliability theory and queuing models to dynamically determine when and how many resources to provision or decommission, using survival functions and cost analysis to balance risk and cost, rather than relying on pre-defined thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If basic thresholding techniques are used for resource allocation, then resource provisioning is simple and fast, but reliability and cost optimization are poor
Solution Approach 1:
The patent transforms the resource allocation approach by changing the parameters from simple threshold values to reliability-based metrics and cost parameters. The system monitors system load alongside reliability metrics and cost data, dynamically adjusting resource allocation decisions based on multiple parameters rather than a single threshold, thereby improving reliability while managing complexity through parameter transformation.
Solution Approach 2:
The patent implements feedback mechanisms where the system continuously monitors resource performance, reliability metrics, and cost data, then uses this feedback to dynamically adjust allocation decisions. The resource allocation system learns from historical data and system responses, creating a closed-loop control system that improves reliability over time while adapting to changing conditions.
2Productivity
If empirical thresholds are used for resource allocation, then implementation is straightforward, but cost modeling and reliability considerations are ignored
Solution Approach 1:
The patent creates a multi-functional resource allocation system that simultaneously handles productivity optimization, cost analysis, and reliability assessment. The system integrates multiple functions into a unified allocation decision process, allowing it to evaluate resource requests based on productivity needs while concurrently considering cost implications and reliability requirements, thus preventing information loss across these different dimensions.
Solution Approach 2:
The patent combines multiple types of information (productivity metrics, cost data, reliability statistics) into a composite decision-making framework. Rather than processing these as separate independent factors, the system integrates them into a unified allocation model that synthesizes diverse information types, similar to how composite materials combine different substances to achieve superior properties.
3Reliability
If automatic scaling is implemented with simple triggers, then response time is fast, but resource optimization and cost effectiveness are reduced
Solution Approach 1:
The patent implements dynamic resource allocation where the system continuously adapts its provisioning decisions based on real-time conditions. Rather than using static trigger-based scaling, the system dynamically evaluates current system state, reliability requirements, and cost parameters to make optimized allocation decisions, allowing it to maintain availability while minimizing unnecessary resource provisioning and associated costs.
Solution Approach 2:
The patent applies partial action by provisioning resources based on actual need rather than over-provisioning for worst-case scenarios. The system calculates the minimum necessary resource allocation to maintain reliability thresholds, avoiding excessive resource provisioning that would increase costs without providing proportional value, thus optimizing the balance between availability and spending.
Data Source
AI summary
Embodiments are directed to a method and system for allocating common resources for a user in a cloud computer network, by: monitoring system reliability and resource charges; estimating a reliability based on the monitored system reliability and simulating resource provisioning choices to measure an impact of resource provisioning on the estimated reliability; estimating a cost based on the monitored resource charges; and allocating network resources based the estimated cost and simulated resource provisioning choices.