Cloud Resource Provisioning via Dynamic Usage Grouping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud services face challenges in efficiently provisioning resources due to unpredictable usage patterns and varying customer needs, leading to inefficiencies in resource allocation and increased costs.
Innovation Solution
Implementing a 'landing pad' resource instance that acts as a universal set of cloud resources for initial provisioning, with dynamic monitoring and grouping of customers based on usage patterns, allowing for the allocation of optimized resource instances as usage evolves.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a universal landing pad resource instance is used for initial provisioning of all computer systems, then resource allocation efficiency is improved and costs are reduced, but the system cannot meet varying customer needs and performance requirements
Solution Approach 1:
The patent segments computer systems into different groups based on their resource usage patterns and characteristics. By analyzing usage data and dividing systems into categories (e.g., high-usage, low-usage, specific-performance-requirements), the system can allocate appropriate resource instances to each group, resolving the contradiction between universal provisioning and customized needs.
Solution Approach 2:
The patent implements dynamic resource allocation where the provisioning system continuously monitors usage patterns and adjusts resource instance assignments in real-time. This dynamic approach allows the system to start with a universal landing pad and evolve to provide customized resource instances as usage patterns emerge, balancing initial efficiency with ongoing adaptability.
2Reliability
If dedicated resource instances are allocated to each computer system to meet specific performance requirements, then customer needs are satisfied, but resource allocation efficiency decreases and costs increase
Solution Approach 1:
The patent applies partial dedication of resources by allowing multiple computer systems to share a resource instance until usage thresholds are exceeded. Instead of fully dedicating resources to each system from the start, the system provides partial exclusivity that can evolve into full dedication only when necessary, maintaining efficiency while ensuring performance when needed.
Solution Approach 2:
The patent changes the allocation parameters dynamically based on monitored usage patterns. Initially, resources are allocated with high sharing parameters for efficiency. As usage patterns emerge and performance requirements are identified, the allocation parameters are adjusted to provide more dedicated resources to specific systems, thus resolving the contradiction between efficiency and performance guarantees.
3Adaptability or versatility
If resource provisioning is customized for each customer from the beginning, then specific needs are met, but the complexity of the provisioning system increases
Solution Approach 1:
The patent implements preliminary provisioning using a universal landing pad resource instance for all computer systems before customization is needed. This preliminary uniform allocation simplifies the initial provisioning process significantly. Later, as usage patterns are monitored and analyzed, customization actions are taken selectively, reducing the overall complexity compared to implementing full customization from the start.
Solution Approach 2:
The provisioning system uses automated monitoring and analysis of usage patterns to self-determine when and how to customize resource allocations. Instead of requiring complex manual configuration for each customer, the system automatically adapts the provisioning strategy based on observed usage, reducing provisioning system complexity while maintaining adaptability.
Data Source
AI summary
A method of provisioning cloud-based applications includes receiving requests from computer systems to provision applications available in a cloud computing environment, assigning each of the computer systems to a first resource instance in the cloud computing environment such that the applications are executed for each of the plurality of computer systems on the first resource instance, monitoring usage of resources for each of the computer systems, grouping each of the computer systems into groups based on the usage of the resources, allocating a second resource instance, and reassigning computer systems in a first group from the first resource instance to the second resource instance.