Improving efficiency and productivity with lakehouse analytics

About

Orange Côte d’Ivoire (OCI), a subsidiary of French telecoms company Orange, is the largest mobile carrier operating in the Ivory Coast with over 17 million active customers. Their services include telecommunications and Orange Money, a banking platform that leverages phone numbers for seamless financial transactions.

Challenge

The main challenge that prompted OCI to reassess their data management strategy was the need for rapid access to data stored in diverse sources. Business needs, particularly in the fight against fraud, necessitated swift access to data.

Additionally, the bottleneck in data extraction requests was becoming increasingly burdensome for the team. Data engineers were spending excessive time fulfilling these requests, leading to delays in delivering insights. Furthermore, they lacked advanced capabilities to modernize their legacy Hadoop data lake, which had performance and governance concerns.

The team initially experimented with open-source Trino, but found it required too much in-house management, and lacked the additional connectivity and security features of Starburst Enterprise.

Solution

Transforming data management with Starburst, Google, and Dell

In response to these challenges, the telecoms company adopted a multi-faceted approach, leveraging the strengths of Starburst, Google Cloud Platform (GCP), and an on-premises Cloud Native platform deployed on Dell Technologies servers. They invested in a new on-premises lakehouse platform, employing Kubernetes and MinIO storage for enhanced performance. They also decided to migrate certain data to Google Cloud Platform (GCP), while keeping sensitive data on-premises to comply with regulatory requirements. The pivotal step in this transformation was the integration of Starburst Enterprise, the fully supported, production-tested and enterprise-grade distribution of open source Trino, which served as the foundation for data virtualization, federation, and the adoption of a data mesh logic.

Partnering with Starburst, OCI provided a thoroughly tested architecture comprising Dell hardware, Starburst Enterprise, and other software on-premises. This collaboration enabled OCI to move data intelligently into a data lakehouse based on usage patterns, preventing vendor lock-in with support for popular open table and file formats. By separating compute and storage and taking advantage of the new technologies implemented,  OCI could create a modern lakehouse at petabyte scale.

Results

The implementation of Starburst yielded significant improvements for OCI:

• Team productivity surged three-fold, thanks to substantial time savings in data analysis requests.
• Business units achieved newfound autonomy through SQL access on Starburst, expediting time-to-insight for data analysts.
• Data Engineers were liberated from repetitive tasks, enabling them to focus on higher value-added projects such as AI development.
• Starburst emerged as the go-to platform for data access, replacing the need to query disparate sources individually.
• Compliance reports that previously took hours to generate on Hadoop now take mere minutes with Starburst.

Looking ahead, Starburst remains a cornerstone in OCI’s data management strategy. It is instrumental in the rollout of data products and the establishment of an open data lakehouse. Through Starburst Stargate, OCI aims to further enhance data federation between on-premises and cloud sources. With plans to integrate Google Big Query with local data, Starburst continues to be a catalyst for OCI’s ongoing journey towards data-driven excellence.