Hierarchical Data Center Capability Summarization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The scalability of cloud computing infrastructure is hindered by the need to manage and process vast amounts of capability information from thousands of backend systems with diverse attributes, leading to potential scalability issues in data storage and computation.
Innovation Solution
A summarization module is deployed across hierarchical levels of the network to cluster nodes based on capabilities, generate histograms, and send summarized data to higher levels, using techniques like k-means clustering and hamming distance for numeric and non-numeric data types, while maintaining relative importance and minimizing data loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all capability information from thousands of backend systems is collected and processed centrally, then complete visibility of device capabilities is achieved, but data storage and computation burdens increase significantly
Solution Approach 1:
The patent divides the data center into hierarchical levels (leaf nodes, intermediate nodes, root nodes) and segments capability information processing across these levels. Each node processes and summarizes capabilities locally before transmitting to higher levels, preventing central overload while maintaining complete capability visibility through distributed hierarchical processing.
Solution Approach 2:
The patent introduces a hierarchical dimension to the capability information architecture, organizing nodes in multiple levels rather than a flat structure. This dimensional change allows capability information to be aggregated and summarized at each level, reducing the burden on any single processing point while preserving complete capability data through the hierarchy.
2Measurement precision
If detailed capability information from all backend systems is transmitted to higher hierarchical levels, then accurate capability representation is achieved, but data transmission volume and processing time increase
Solution Approach 1:
The patent extracts only the essential capability information needed at each hierarchical level, rather than transmitting all raw capability data. Summarization logic at each node identifies and transmits only the most relevant capability characteristics to higher levels, maintaining accurate capability representation while significantly reducing data transmission volume and processing time.
Solution Approach 2:
The patent applies partial action by transmitting summarized capability information rather than complete raw data at each hierarchical level. This partial transmission is sufficient for higher-level decision-making and resource allocation, avoiding the excessive time and bandwidth costs of transmitting all detailed capability information while maintaining adequate representation accuracy.
3Adaptability or versatility
If capability information from diverse backend systems with different attributes is aggregated, then comprehensive capability coverage is achieved, but data heterogeneity and processing complexity increase
Solution Approach 1:
The patent creates a universal capability information framework that can handle diverse backend system attributes through a common hierarchical structure. The summarization logic at each node adapts to handle different data types and attributes uniformly, enabling comprehensive capability coverage across heterogeneous systems while managing data diversity through a standardized multi-functional processing approach.
Solution Approach 2:
The patent transforms heterogeneous capability parameters from diverse backend systems into a standardized hierarchical representation. By changing the parameter structure from raw diverse attributes to summarized hierarchical capability descriptors, the system achieves comprehensive capability coverage while reducing the complexity of managing data heterogeneity through parameter transformation and normalization.
Data Source
AI summary
A method for summarizing capabilities in a hierarchically arranged data center includes receiving capabilities information, wherein the capabilities information is representative of capabilities of respective nodes at a first hierarchical level in the hierarchically arranged data center, clustering nodes based on groups of capabilities information, generating a histogram that represents individual node clusters, and sending the histogram to a next higher level in the hierarchically arranged data center. Relative rankings of capabilities may be used to order a sequence of clustering operations.