Cloud Resource Planning via Usage Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing resource providers face challenges in efficiently managing and allocating computing resources due to varying customer demands, leading to inefficiencies in resource utilization and potential overprovisioning or underprovisioning of resources.
Innovation Solution
A resource planning application that collects usage statistics, classifies customer usage patterns, and forecasts demand to optimize the allocation and procurement of computing resources, ensuring that resources are adequately provisioned to meet customer needs while minimizing waste and costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If computing resources are reserved for all customers based on promised allocation, then customer service reliability is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent implements dynamic resource allocation that adjusts resource provisioning based on actual customer usage patterns and demand forecasts. Instead of static reserved allocation, the system continuously monitors usage statistics and modifies resource allocation in real-time, allowing resources to be dynamically assigned to customers who need them while being released from idle allocations, thus resolving the contradiction between reliability and utilization efficiency
Solution Approach 2:
The system performs preliminary demand forecasting and resource planning based on historical usage patterns before actual resource allocation. By predicting future customer needs in advance, the system can pre-provision resources efficiently, ensuring customer service reliability is maintained while avoiding over-provisioning of resources that would reduce utilization efficiency
2Reliability
If additional data storage capacity is installed to meet peak demand, then customer service reliability is improved, but infrastructure cost increases
Solution Approach 1:
The patent changes the parameter of resource allocation from fixed physical infrastructure to flexible virtualized resource allocation. By using virtualization and dynamic provisioning, the system can meet peak demand reliability requirements without permanently installing additional physical storage capacity. Resources are allocated based on changing parameters of actual customer needs rather than fixed infrastructure over-provisioning
3Reliability
If resource allocation is increased to accommodate unstable usage patterns, then service level agreement compliance is improved, but resource waste increases
Solution Approach 1:
The system implements continuous feedback loops that monitor actual customer usage patterns and compare them against service level agreements. This feedback enables the system to distinguish between stable and unstable usage patterns and adjust resource allocation accordingly, ensuring SLA compliance for customers with unpredictable needs while preventing resource waste by not over-provisioning for customers with stable, predictable usage patterns
Data Source
AI summary
Disclosed are various embodiments for the planning of resources used in computing. Usage statistics regarding one or more virtual machine instances executing in a networked plurality of computing devices are obtained. The usage statistics are grouped, for example, based on one or more customer usage classifications, thereby producing one or more usage groups. A corresponding demand forecast is generated for each of the usage groups. A projected demand for one or more physical components of the networked computing devices is calculated according to the demand forecasts.