←←All cases

Data warehouse modernization for telecom giant

Client

Multinational telecommunications company

Industry

Telecommunications

Services

Software Development, Data Engineering, Data Architecture, Data Warehousing, Cloud Migration

Tech

Microsoft Fabric, OneLake, Power BI, PySpark, Microsoft Purview, Azure Data Lake Storage (ADLS Gen2)

Challenge

Serving over 15 million subscribers and processing billions of call detail records (CDRs) daily, an EU-based telecommunications leader had reached the point where its infrastructure was actively working against its business. Executives were making daily commercial decisions based on yesterday's data because the on-premises Oracle database beneath it all was collapsing under its own load. The numbers tell the story: over 30,000 fragmented reporting flows, many of which were running outdated scripts that frequently failed and delayed daily revenue insights. Extended processing times and recurring script failures had become commonplace. The constraint was not only technical, but also financial: the client needed to expand its data capabilities without increasing its budget. The company needed a partner who could provide expertise in modernizing data warehouses and approached ITRex for exactly that.

Assigned task

●

Initially, the request was clear-cut: integrate five ITRex data engineers into the in-house setup to bolster data warehousing operations

●

The initial scope was augmentation, but the team quickly went deeper—learning the client's full data landscape, from CDR processing intricacies to subscriber analytics

●

Given the sensitivity of telecom data, strict adherence to the client's data security protocols was non-negotiable—every CDR and subscriber detail had to stay confidential

●

As an external team, fitting into the client's organizational culture and working rhythm was essential to a cohesive partnership

●

As ITRex demonstrated capability and reliability, the client expanded the engagement—from data engineering support to leading full project rollouts

ITRex kicked off the engagement by stabilizing core ETL operations in the on-site Oracle database—pooling data from call centers, retail systems, and external providers to protect daily revenue insights and business continuity while the bigger modernization work got underway. From there, the team moved systematically:

●

Integration of new data sources. Connected and integrated numerous new data sources into the existing system, expanding the platform's reach without destabilizing it.

●

Reporting cleanup. Managed 30,000+ report workflows, auditing and removing outdated ones to establish a lean, maintainable system.

●

Direct data access for analysts. Gave analysts direct access to raw data layers, cutting report generation cycles from days to under an hour.

●

Team growth. Diversified the ITRex team with additional data engineers, testers, business analysts, and project managers to support the rollout of new reporting processes—growing from the original five-person augmentation to a 50-member dedicated team over the course of a 15+ year partnership.

●

Data warehouse modernization. Added new layers to the data warehouse and built data marts for advanced scaling and analysis.

●

Safe testing environment. Set up a secure environment to test new reporting functionality against near-live data, with zero risk of exposure.

●

Performance optimization. Resolved script-level glitches in the legacy reporting tools and optimized processing performance throughout.

●

Training & knowledge transfer. Delivered extensive analyst training to support smooth adoption of the new system.

Cloud migration to Microsoft Fabric

ITRex led the migration of the client's data platform to Microsoft Fabric, replacing the fragmented on-premises Oracle environment with a unified, cloud-first architecture built around the following components:

●

Unified storage via OneLake & a Medallion architecture (Bronze/Silver/Gold). The multi-tenant Oracle environment was consolidated into a single OneLake tenant, with billions of daily CDRs, retail transactions, and call center logs ingested into a unified lakehouse stored in open-source Delta Parquet format—ensuring ACID compliance and eliminating proprietary data silos.

●

Serverless compute & PySpark refactoring. Over 30,000 legacy Bash crons and PL/SQL scripts were refactored into parameterized PySpark notebooks running on Fabric Synapse Data Engineering. Fabric's instant-on Spark pools eliminated cluster spin-up delays, enabling memory-intensive deduplication of billions of CDR rows to run concurrently.

●

OneLake Shortcuts for zero-copy integration. During the hybrid migration phase, OneLake Shortcuts linked directly to existing ADLS Gen2 accounts, letting Fabric compute engines query legacy assets without duplicating data.

●

Direct Lake mode for Power BI. Power BI was deployed using Fabric's native Direct Lake mode, reading Delta Parquet files straight from OneLake without intermediate SQL queries or an Analysis Services cache—giving business analysts real-time visibility into multi-terabyte datasets without performance degradation.

●

Governance via Microsoft Purview. To meet strict European telecom regulations, data security was built in natively using Microsoft Purview, with row-level and column-level security applied directly at the OneLake level to keep subscriber PII protected across every downstream view.

data warehouse modernization for telecom giant

Impact

The modernized platform changed what "fast" meant for this client—and the gains showed up across operations, cost, and the business itself.

●

A 3× increase in CDR processing and analytical reporting speed, turning hours-long database queries into automated routines completed in minutes

●

A 34% reduction in IT infrastructure and operational costs within the first 14 months, with a 240% ROI by the end of year one—driven by cloud migration, tool consolidation, and the elimination of redundant licensing and on-premises hardware

●

Time to value dropped from 6 months to 3 weeks for the first operational dashboards, a direct result of removing the manual reporting bottlenecks that had defined the legacy system

●

A 15% reduction in subscriber churn, enabled by near-real-time CDR availability that lets marketing and customer success teams launch targeted retention campaigns the same day

●

25% faster time-to-market for personalized telecom offerings and data plans, made possible by giving analysts immediate access to the data layers needed for agile behavioral and trend analysis

●

An estimated 8% increase in Average Revenue Per User (ARPU), driven by sharper subscriber segmentation built on the unified data warehouse