Building a large data warehouse in order to bill clients and provide a B2B activity dashboard
Optimizing global portal databases helping retailers and cities to monitor the recycling of 50 billion used bottles and cans a year in the world
Epay
CUSTOMER SPOTLIGHT
Epay focuses on providing comprehensive financial services such as online deposit and withdrawal services, online payment, currency exchange, and global remittance. Epay has a total of 1,000,000 global users and more than 200 cooperative banks and financial institutions, connecting with all major financial regions of the world. Epay leads financial services with "technological innovation", and is committed to becoming the most popular payment system in the world. Epay wishes to build a cross-border payment clearing center, link with the global payment channels and merchants, and lead the global payment innovation development.
Project Challenge
Epay had issues invoicing their clients without a data warehouse, based on data processing activities all across the globe. Because of the delay, Epay had delaying invoicing for up to 60 days and cash management was an issue. Epay had set up a multi-tenancy database system with servers located across the globe for data privacy compliance. The databases were carrying the data generated by their software INGEVOUCHER with different firmware upgrades per client. In some cases, the data model was slightly different as the upgrades were delayed based on different maintenance windows.
Solution
Phase 1: The Dwh
CHDS implemented the data warehouse solution with all the data pipelines to extract the data from the operation stores. The particularity of the data pipelines were to handle dynamically the meta data in order to manage schema augmentation based on feld and data type names. This clever data pipeline set up enabled Epay to extract heterogeneous datasets into a common data warehouse.
Phase 2: The Billing system
Epay billing is very close to the billing system of telecommunication companies with a rating grid: the bigger the activity the lower the cost for the client it is. CHDS implemented a rating mechanism to calculate the actual activity volume and the actual rate via a grid of rates based on volume per product. Once the rating per month was done, billing was able to be calculated and invoiced, removing delays. Epay was then able to invoice clients efficiently, gaining 45 days of cash flow.
Phase 3: The Dashboard
Clients lacked visibility for their platform metrics. In order to fix this issue, a MOLAP cube was built to gather information from the enterprise data warehouse. When the client connected to the dashboard, a session was created with a subcube, to isolate the client data from any other client data, with a method called slicing on the client dimension. A master database was also created to manage the different widgets allowed in the dashboard to be dragged and dropped onto the GUI for the client. Each widget performed multidimensional expression (MDX) onto the cube, ensuring a linear performance and a great scalability on an OLAP storage with warmed pre-aggregations.