Cloud Resource Allocation via Global-Local Iterative Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Efficient allocation of computing resources in cloud computing environments is challenging due to their highly distributed nature, making it difficult to accurately determine the optimal location for resource placement.
Innovation Solution
A two-level resource allocator system comprising a global allocator and local allocators, where the global allocator generates estimates for resource placement based on general information and iteratively refines the location using detailed information from local allocators until an appropriate placement is determined.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a global allocator is used to manage distributed cloud computing resources, then resource allocation coverage is improved, but allocation accuracy deteriorates due to lack of local detailed information
Solution Approach 1:
The system divides the cloud computing environment into multiple local portions, each managed by a local allocator that has detailed knowledge of its local resources. The global allocator is segmented into a coordinator that manages overall resource distribution. This segmentation allows each allocator to operate within its expertise domain, resolving the contradiction between global coverage and local accuracy.
Solution Approach 2:
The patent implements a nested allocator structure where local allocators are embedded within the global allocator framework. The global allocator contains multiple local allocators, each nested within specific geographic or logical regions. This nested architecture enables the global system to maintain both broad coverage and localized precision simultaneously.
2Measurement precision
If local allocators are introduced to improve allocation accuracy, then system complexity increases due to multiple allocation layers
Solution Approach 1:
Both global and local allocators implement the same core allocation logic and interface protocols, making them universally compatible. The local allocators reuse the proven allocation algorithms from the global allocator, adapting them to local contexts. This multi-functionality reduces system complexity by eliminating the need for entirely different allocation mechanisms at each level.
Solution Approach 2:
The global allocator acts as an intermediary between resource requests and local allocators, coordinating their activities and consolidating their outputs. This intermediary role simplifies the interaction between multiple allocators, providing a unified interface that masks the underlying complexity of having multiple allocation layers.
3Measurement precision
If iterative refinement between global and local allocators is implemented, then resource placement accuracy is improved, but allocation time increases
Solution Approach 1:
Local allocators pre-process and cache information about their local resource availability and characteristics before global allocation decisions are made. This preliminary action allows the global allocator to make faster, more informed decisions without requiring extensive iterative refinement, thus reducing allocation time while maintaining accuracy.
Solution Approach 2:
The system implements feedback loops where local allocators provide real-time information about resource status and allocation outcomes to the global allocator. This feedback mechanism enables the global allocator to refine its decisions based on actual local conditions, improving placement accuracy without requiring multiple iterative cycles, as the feedback provides actionable insights in real-time.
Data Source
AI summary
A device, of a cloud computing environment, receives information associated with one or more computing resources of a local portion of the cloud computing environment, and receives, from a global allocator device, a request for placement of a computing resource element at a selected computing resource of the one or more computing resources. The device also determines, based on the information, whether the request can be satisfied by the selected computing resource, and places the requested computing resource element in the selected computing resource when the request can be satisfied by the selected computing resource. The device sends, to the global allocator device, a message indicating that the requested computing resource element has been successfully placed in the selected computing resource.