Disaggregated Resource Map Construction for Coherency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to effectively address the challenges of building and accessing disaggregated resources due to issues with unreliable networks, resource consumer and node independence, resource coherency, and fine-level access, particularly in distributed computing environments where resources are decentralized and operate without a centralized controller.
Innovation Solution
The solution involves resource nodes providing information to consumers, allowing them to create detailed maps of the disaggregated resources, ensuring fine-level access and coherency, with resource consumers constructing their own views and accessing resources independently without relying on extraneous systems, and using split resource maps for granular access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If resources are disaggregated into individual resource nodes without a centralized controller, then scalability and cost are improved, but network reliability and resource coherency deteriorate due to inherent network unreliability and latency
Solution Approach 1:
The patent segments the resource map into multiple distributed copies stored across different resource nodes. Each resource node maintains a complete or partial copy of the resource map, eliminating the need for a centralized controller while ensuring that any node can provide the necessary mapping information, thus maintaining reliability through distribution.
Solution Approach 2:
The patent implements feedback mechanisms where resource nodes continuously exchange information about resource availability and map consistency. This allows the distributed system to detect and correct coherency issues, ensuring that all nodes maintain synchronized views of the resource state despite network latency and potential failures.
2Productivity
If resource consumers construct their own views of disaggregated resources, then independence and responsiveness are improved, but the complexity of managing multiple different views deteriorates
Solution Approach 1:
The patent employs copying by having each resource consumer maintain a local copy of the resource map. This local copy enables consumers to access resource information independently without querying multiple nodes, thus improving responsiveness while the standardized copy format reduces the complexity of managing different views.
Solution Approach 2:
The patent enforces homogeneity by standardizing the resource map structure and format across all resource nodes and consumers. This uniform structure allows any consumer to interpret the map consistently, reducing the complexity that would otherwise arise from managing heterogeneous views of the disaggregated resources.
3Measurement precision
If fine-level granularity access is implemented below resource device level, then access precision is improved, but the difficulty of maintaining complete and detailed maps deteriorates
Solution Approach 1:
The patent introduces an additional dimensional layer to the resource map structure, organizing fine-grained access information in a hierarchical or multi-dimensional format. This allows detailed mapping information to be structured in a way that is easier to manage and verify, reducing the difficulty of maintaining complete maps while preserving fine-level access precision.
4Productivity
If resource nodes operate independently without out-of-band communication, then performance and simplicity are improved, but the ability to maintain coherency across nodes deteriorates
Solution Approach 1:
The patent ensures continuity of useful action by designing the resource map copying and synchronization mechanisms to operate continuously in the background without interrupting the primary resource access operations. This allows nodes to maintain coherency through continuous information exchange while preserving high performance for actual resource operations.
Data Source
AI summary
Disaggregated resources distributed among resource nodes provide access to resource consumers by offering resource node information to the resource consumers. Resource node information supplied by each individual resource node comprises incomplete information with respect to the complete disaggregated resource. Resource consumers collect resource node information to create maps of the disaggregated resource, ensure coherency, or manage the disaggregated resource.


