Telco giant reduces costs managing unstructured data
A telecommunications provider needed to establish best practices for the use of NoSQL and open-source databases that would result in cost savings. Pythian delivered the solution.
What we did
- Tested the feasibility of replacing Oracle for a specific time-series use case: aggregating billing events to generate invoices at a given cadence (30 to 60 days); to perform the test, Pythian delivered a time-series Cassandra PoC hosted on Google Cloud virtual machine instances
- Selected Terraform to provision the virtual machines, and Puppet automated the installation of Apache Cassandra
- Exported client’s billing events from their Oracle database as a CSV file, which was ingested via Apache Spark
- Google Cloud Platform
- Apache Cassandra