Subscriber Data Point Redundancy With Flag-Based Job Coordination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the 5G network charging system, the subscriber data point (SDP) system is not designed for complete redundancy, leading to potential loss of up-to-date subscriber data when a server is down, causing inaccuracies in billing and difficulty in data recovery.
Innovation Solution
Implementing a method where two or more servers are scheduled to perform the same jobs, with one server setting a flag indicating job completion, and the other server checks the flag after a preset time to avoid duplication, ensuring availability of up-to-date subscriber data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If subscriber data is replicated to two or more partitions with multiple servers scheduled to run jobs, then system availability is improved, but certain jobs (bill cycle, life cycle, subscriber database snapshot) are only run by one partition, causing data loss when that server is down
Solution Approach 1:
The system performs preliminary actions by scheduling the same critical jobs (bill cycle, life cycle, subscriber database snapshot) on multiple servers in advance. Each server is pre-configured with the capability to execute these jobs, so when one server is running them, other servers are ready to take over immediately. This preliminary preparation ensures that job execution capability is maintained even when individual servers fail.
Solution Approach 2:
The system implements beforehand cushioning through comprehensive data replication and job scheduling redundancy. Subscriber data is replicated across multiple partitions, and multiple servers are scheduled to run the same critical jobs. This creates a buffer or cushion against server failures, ensuring that when one server goes down, the replicated data and scheduled jobs on other servers prevent information loss and maintain system reliability.
2Reliability
If multiple servers are scheduled to run the same jobs, then data redundancy is improved, but duplicate job execution may occur without coordination
Solution Approach 1:
The system implements feedback mechanisms where servers monitor and communicate the status of critical jobs. When one server successfully executes a job (such as bill cycle, life cycle, or subscriber database snapshot), this information is fed back to other servers scheduled to run the same job. This feedback allows the system to prevent duplicate execution by informing other servers that the job has already been completed, thereby maintaining data redundancy while avoiding unnecessary duplicate operations.
Solution Approach 2:
The system introduces an intermediary coordination mechanism that manages job execution across multiple servers. This intermediary layer coordinates between servers running the same critical jobs, ensuring that when one server executes a job, other servers are notified or blocked from executing it again. This mediator approach simplifies the complexity of direct server-to-server coordination while maintaining proper job execution control and data redundancy.
3Device complexity
If only one partition runs critical jobs, then system complexity is reduced, but billing accuracy and data recovery capability deteriorate when the server is down
Solution Approach 1:
The system applies local quality by making different parts of the system (different server partitions) have different roles and capabilities. Specifically, multiple partitions are configured with the local quality to execute critical jobs (bill cycle, life cycle, subscriber database snapshot), rather than concentrating this capability in a single partition. This distributed local quality ensures that billing accuracy and data recovery capability are maintained even when one server fails, as other partitions with the same job execution capability can take over.
Data Source
AI summary
Methods and systems to achieve complete redundancy for a subscriber data point are disclosed. The subscriber data point (SDP) may be implemented on an account balance management function (AMBF) of a 5G charging system. According to an implementation, a first computing device of the AMBF may be scheduled to perform an operation on subscriber data stored in a first database, and a second computing device of the AMBF may be scheduled to perform the same operation on subscriber data stored in a second database after a preset time period from the first computing device performing the operation. The first computing device may set a flag in the first database and the second database when the operation starts. The second computing device may check the flag status when starting the operation. Upon detecting the flag is on in the second database, the second computing device may terminate the operation.