Bandwidth Efficient KPI Filtering in Distributed Databases
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for measuring Key Performance Indicators (KPIs) in distributed networks are overly complex and inefficient, consuming large network resources and not scalable, especially in distributed network topologies where data is forwarded to a central location for processing.
Innovation Solution
A method that reduces data movement between edge nodes and a central querying node by using an iterative approach with two or more queries, where the central node sends queries to edge nodes to cache subscriber data and return filtered results, followed by targeted subsequent queries to aggregate data efficiently, conserving processing and bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all subscriber data is forwarded to a central location for grouping, sorting, and processing, then complete KPI measurement is achieved, but network bandwidth consumption and processing complexity increase significantly
Solution Approach 1:
The patent divides the centralized data processing task into distributed segments performed at edge nodes. Each edge node independently performs local data filtering and aggregation, returning only necessary results to the central node. This segmentation reduces network bandwidth consumption while maintaining measurement completeness by distributing the computational workload across multiple locations in the network hierarchy.
Solution Approach 2:
The patent applies preliminary filtering and aggregation actions at edge nodes before data reaches the central location. By pre-processing data locally and returning only essential results, the system reduces the volume of data requiring central processing, thereby decreasing network bandwidth consumption and central processing complexity while preserving complete KPI measurement capability.
2Measurement precision
If all subscriber data is forwarded to a central location for processing, then complete KPI measurement is achieved, but system complexity and resource requirements increase
Solution Approach 1:
The patent segments the processing system into distributed edge nodes that perform local filtering and aggregation operations. This segmentation reduces central processing complexity by distributing computational tasks, while each edge node's simplified local processing maintains overall measurement completeness without requiring complex centralized coordination.
Solution Approach 2:
Edge nodes autonomously perform data filtering, aggregation, and preliminary processing without requiring complex central coordination. This self-service capability at the edge reduces the complexity of the central processing system while maintaining complete KPI measurement through distributed intelligent processing.
3Loss of energy
If iterative querying is used to reduce data movement, then bandwidth efficiency improves, but query complexity and processing time increase
Solution Approach 1:
The patent performs preliminary filtering and aggregation at edge nodes before central collection, pre-processing data to minimize subsequent query requirements. This preliminary action reduces the amount of data requiring iterative querying and central processing, thereby improving bandwidth efficiency without significantly increasing overall processing time.
Solution Approach 2:
The patent implements partial data retrieval through iterative querying, where edge nodes return only the most relevant results in successive queries rather than all data. This partial action approach optimizes bandwidth efficiency by transmitting only essential information, while the iterative nature manages to balance query complexity with processing time requirements.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A central node sends a query indicating at least one key performance indicator (KPI) and a first KPI filter criterion to one or more edge nodes causing each edge node to cache subscriber data from each subscriber having a KPI value corresponding to the KPI. The central node receives subscriber data, responsive to the query, and aggregates the subscriber data according to each subscriber and the corresponding KPI value to yield an aggregated initial subscriber dataset and sends a subsequent query to each edge node that (for each edge node) indicates a subscriber from the initial subscriber dataset not returned by the edge node. The central node receives subsequent subscriber, responsive to the subsequent query, and aggregates the subsequent subscriber data with the initial subscriber dataset to yield a final subscriber dataset that indicates a total KPI value for each subscriber.