Call Data Block Management for Near Real-Time Network Fault Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional database storage methods for call data in cellular communication networks are inefficient, leading to slow access times and inability to handle large volumes of data in near real-time, especially when accessing billions of records, which hinders quick fault detection and analysis in wireless communication systems.
Innovation Solution
A method of managing call data by assembling call data records into data blocks and storing them in a way that allows for per-block retrieval, along with spatial and temporal grouping, and indexing using single dimensional coordinates, enabling faster access and processing of large datasets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If conventional database storage methods are used to store call records, then each record can be read individually and independently, but accessing large numbers of records requires multiple slow disk searches and retrieval is prohibitively delayed
Solution Approach 1:
Multiple call data records are merged into a single composite data structure called a 'super-record' that contains multiple individual call records. This allows the system to retrieve multiple records in a single disk access operation rather than requiring separate disk searches for each record, thereby dramatically reducing access time and improving retrieval efficiency when querying large numbers of records.
Solution Approach 2:
Call data records are pre-organized and grouped into super-records during the data ingestion phase, rather than storing them as individual isolated records. This preliminary structuring enables faster query performance later, as the system can directly access pre-grouped records without performing multiple sequential disk searches.
2Ease of operation
If individual call records are stored separately in database, then records can be updated or refreshed independently, but searching through billions of records to access relevant data is prohibitively slow
Solution Approach 1:
The system segments call data into two levels: individual call records that can be updated independently, and composite super-records that group multiple individual records together for efficient retrieval. This segmentation allows the system to maintain the independence of individual record updates while enabling fast batch retrieval through the super-record structure, resolving the contradiction between update flexibility and search speed.
3Productivity
If traditional database storage is used, then only a few records need to be accessed at a given point in time with high efficiency, but when billions of records are accessed the delay is prohibitively long preventing near real-time analysis
Solution Approach 1:
The patent combines multiple individual call records into composite super-records that can be retrieved in a single disk access operation. This merging enables the system to return large volumes of data (potentially billions of records) much faster, reducing the delay between query generation and data return from hours or days to near real-time, thereby enabling continuous data analysis.
Data Source
Figure 1
Figure 2~3
Figure 4~6
AI summary
A method of managing call data for at least one radio network element within a cellular communication network. The method comprises receiving call data for at least one call from the at least one radio network element within the cellular communication network, arranging the received call data into call data records, assembling the call data records into at least one data block, and writing the at least one data block to at least one data storage device. The method further comprises, upon receipt of a call data query, retrieving call data records from the at least one data storage device on a per data block basis.