The billing system processes a large amount of continuously incoming data. Tasks – are account for the movement of funds on customer accounts – calls, subscriptions, charges, payments and adjustments. There is accounting of services provided to subscribers.
There were bottlenecks in billing. Data collection between data centres took up a significant share of traffic. This limited the number of new events accepted. At the close of the month, billing took a working day and required the involvement of engineers in the process.
The new system works reliably, and easily copes with loads. the solution is to speed up data saving, NoSQL NOS. This relieved the network traffic of the system responsible for receiving it. As a result, we increased the load threshold by events from 0.5 billion to 2 billion per day.
We implemented MapReduce-algorithms, reduced the closing time of the month from ~1 day to ~10 minutes, automated parameter setting. Now the system does not require the participation of engineers to perform scheduled operations.
We implemented a mechanism for faster deployment to the customer’s environment. Big big data / fast data meets the requirements for fault tolerance and scaling.
We used the following technologies:
The platform is a reliable tool for rapid visual development of complex IT systems. On the platform you can create and develop:
The main thing in solutions is the speed of development and flexibility. The platform guarantees a high level of security, stability and scalability.
Digital BSS operator system is a platform for effective digital transformation. There was a need to improve this solution to a boxed product.
Initially, the system was a large-scale multi-module solution with numerous APIs. The disadvantage of this implementation was the complexity of configuring the system for a specific client.
To simplify API support for various clients, GraphQL technology was introduced into the platform. API of a set of Backend modules collected under one proxy module. The proxy module provides a complete and clear description of the platform API data model, as well as access to this data, for the integration of client Frontend systems.
When integrating with the platform, customers were given the opportunity to request what they need, while minimizing the work on changing the platform itself.
The platform now provides an API aimed at boxed delivery. In addition, for the possibility of customization, we have divided the configuration of the GrapthQL API platform model into basic and client. Thanks to this, customers can adapt the platform to their requirements.
The project used technology: