Data Locality Estimation via Cable Length Grouping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Software-defined infrastructures face challenges in maintaining data availability and locality due to the lack of granular awareness of node physical locations, leading to potential single points of failure when resources are shared across nodes.
Innovation Solution
A method is introduced to estimate physical disparity for data locality by obtaining vital product data from cables connecting nodes to switches, determining cable lengths, assigning group identifiers based on these lengths, and selecting nodes for data storage to mitigate the risk of resource failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If nodes are grouped by cable length to improve data locality awareness, then data availability is improved, but system complexity increases due to additional grouping and tracking mechanisms
Solution Approach 1:
The patent introduces an intermediary management application that acts as a mediator between the physical infrastructure (nodes, cables, switches) and the data storage operations. This application automatically collects cable length information, assigns group identifiers to nodes, and manages data placement decisions, thereby resolving the contradiction by automating the complexity-handling functions rather than requiring manual tracking or complex distributed algorithms.
Solution Approach 2:
The system performs preliminary actions by pre-collecting cable length information from Vital Product Data (VPD) during system initialization or node addition, and pre-assigning group identifiers to nodes based on their physical characteristics before data storage decisions are made. This advance preparation eliminates the need for complex real-time calculations when placing data, simplifying the operational complexity while maintaining reliability.
2Measurement precision
If cable length information is collected from VPD to determine physical location, then measurement precision is improved, but ease of operation deteriorates due to additional data collection and processing requirements
Solution Approach 1:
The patent implements self-service by having the management application automatically collect cable length information directly from the Vital Product Data (VPD) stored in the cable or node hardware itself. The system queries the VPD through standard interfaces, parses the length information, and uses it autonomously for grouping decisions without requiring manual measurement or configuration by operators, thereby maintaining ease of operation while achieving precise physical location awareness.
3Loss of information
If nodes are assigned group identifiers based on cable length, then data locality awareness is improved, but device complexity increases due to additional tracking and management requirements
Solution Approach 1:
The patent applies parameter changes by transforming the continuous physical parameter of cable length into a discrete group identifier parameter. Instead of tracking exact cable lengths for all nodes, the system categorizes nodes into groups based on cable length ranges, thereby reducing the information management complexity from continuous measurements to discrete categories while preserving sufficient locality awareness for effective data placement decisions.
Data Source
AI summary
Methods, apparatuses, and computer program products for estimating physical disparity for data locality in software-defined infrastructures are disclosed. For each node in a cluster of nodes connected to a switch, vital product data (VPD) of a cable connecting the node to the switch is obtained, and for each cable, a length of the cable is determined from the VPD. A management application assigns a group identifier to each node in the cluster based on the length of the cable connecting the node to the switch. The management application selects a node in the cluster for storing a data set in dependence upon the group identifier of the node.