Hierarchical Cloud Resource Allocation with Layered Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for provisioning cloud-based services are inefficient in terms of time and resource usage, particularly in complex systems requiring multiple resources, as they either fail due to incomplete resource availability, result in rigid architectures, or take excessively long to identify optimal solutions.
Innovation Solution
A method that identifies and allocates resources by generating a hierarchical plan, allowing each architectural layer to communicate only with its immediate neighbors, selecting eligible resources, and recursively requesting resources from lower layers, ensuring efficient allocation without architectural rigidity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the brute-force approach is used to allocate resources, then the allocation process is simple and straightforward, but it fails when not all required resources are available and requires undoing all prior allocations, making it time-consuming and inefficient
Solution Approach 1:
The patent performs preliminary identification of all required resources and their availability before attempting allocation. By pre-checking resource availability and identifying eligible resources in advance, the system可以避免 brute-force allocation failures and eliminates the need to undo prior allocations, thus resolving the contradiction between ease of allocation and allocation success rate.
2Reliability
If the optimal solution approach is used to identify resources, then all possible solutions are evaluated to find the best one, but it can take an infinitely long time for large systems, making it time-consuming
Solution Approach 1:
The patent segments the resource identification process into hierarchical layers, where each layer independently identifies and selects eligible resources. This segmentation allows parallel processing and eliminates the need to evaluate all possible solutions exhaustively, thus reducing identification time while maintaining reliable resource selection through layered evaluation.
3Productivity
If the architectural layers are merged to gain visibility into all aspects of service topology, then resource allocation becomes efficient, but the architecture becomes rigid and inflexible, requiring changes to affect all other layers
Solution Approach 1:
The patent maintains separate architectural layers with distinct responsibilities, where each layer independently identifies and selects resources. This segmentation preserves architectural flexibility and adaptability while achieving efficient resource allocation through coordinated communication between layers, thus resolving the contradiction between allocation efficiency and architectural flexibility.
4Ease of manufacture
If multiple resources are allocated in sequence using brute-force, then the allocation process is simple, but if a resource is not available, all prior allocations must be undone, resulting in inefficient resource usage
Solution Approach 1:
The patent performs preliminary identification of all required resources and checks their availability before attempting allocation. By pre-checking resource availability and identifying eligible resources in advance, the system可以避免 brute-force allocation failures and eliminates the need to undo prior allocations, thus resolving the contradiction between ease of allocation and allocation success rate.
Data Source
AI summary
Systems, methods, and computer readable media for identifying resources to implement a service in a cloud computing environment are disclosed. In general, the disclosed methodologies analyze a cloud's ability to support a desired service while maintaining separation between the cloud's logical layers. For example, given a list of resources needed to implement a target service, a hierarchical plan may be generated. The plan may then be used by each layer to track and record the availability of various possible layer-specific resource selections. Since each layer may be permitted access only to that portion of the plan that is associated with, or applicable to, the specific layer, the logical separation between different layers may be enforced. As a consequence, each layer may implement its resource selection mechanisms in any desired manner.