Databricks > Case Studies > AT&T's Cloud Migration for Enhanced Customer Service and Operational Efficiency

AT&T's Cloud Migration for Enhanced Customer Service and Operational Efficiency

Technology Category

Analytics & Modeling - Machine Learning
Platform as a Service (PaaS) - Application Development Platforms

Applicable Industries

Cement
Telecommunications

Applicable Functions

Logistics & Transportation
Product Research & Development

Use Cases

Construction Management
Time Sensitive Networking

Services

Cloud Planning, Design & Implementation Services
Data Science Services

About The Customer

AT&T is a technology giant and one of the leading communication service providers in the world. The company has hundreds of millions of subscribers and ingests over 10 petabytes of data across its entire data platform each day. To harness this data, AT&T has a team of over 2,500 data users across more than 60 business units. The company's business is data-powered, from building analytics to ensure decisions are based on the best data-driven situation awareness to building machine learning models that bring new innovations to its customers. AT&T's goal is to increase its use of insights for improving its customer experience and operating more efficiently.

The Challenge

AT&T, a leading telecommunications company, was grappling with the challenges of managing a massive on-premises legacy Hadoop system. The system was complex and costly, hindering operational agility, efficiency, and the optimal use of engineering resources. AT&T needed to transition to the cloud to better support its hundreds of millions of subscribers. The company was dealing with a highly complex hardware setup, including 12,500 data sources and over 1,500 servers. The on-premises Hadoop architecture was not only expensive to manage but also complex to maintain. Operational costs were high, and additional capital costs were incurred for data centers, licensing, and more. Up to 70% of the on-prem platform had to be prioritized to ensure 50K data pipeline jobs met SLAs and data quality objectives. Engineers were spending more time managing updates, fixing performance issues, and provisioning resources rather than focusing on higher-value tasks. The physical infrastructure's resource constraints also slowed down data science activities, impeding innovation.

The Solution

AT&T decided to migrate from Hadoop to Databricks on the Azure cloud. The company conducted a comprehensive cost analysis and concluded that Databricks was the fastest and most cost-effective solution for data pipelines and machine learning workloads. The migration process was meticulously planned, with the largest workloads being migrated first to immediately reduce the infrastructure footprint. The data was migrated before the users to ensure a smooth transition for the thousands of data practitioners. The company spent a year deduplicating and synchronizing data to the cloud before migrating any users. This was a crucial step in successfully migrating a large, complex multi-tenant environment of over 2,500 users from more than 60 business units and their workloads. The user migration process took nine months, allowing AT&T to retire on-premises hardware in parallel with the migration to accelerate savings. The entire migration process took about two and a half years and was managed with the equivalent of 15 full-time internal resources.

Operational Impact

The migration to Databricks on the Azure cloud has brought significant benefits to AT&T. The company has experienced substantial cost savings and has been able to rationalize about 30% of its data by identifying and not migrating underutilized and duplicate data. The migration of the largest workloads first allowed half the on-prem equipment to be rationalized during the migration. The result is an anticipated 300% five-year migration ROI from OpEx savings and cost avoidance. With data readily available and the means to analyze data at any scale, teams of citizen data scientists and analysts can now spend more time innovating. Data scientists are now able to collaborate more effectively and speed up machine learning workflows so that teams can deliver value more quickly. AT&T now has a single version of truth to create new data-driven opportunities, including a self-serve AI-as-a-Service analytics platform that will enable new revenue streams and help it continue delivering exceptional innovations to its millions of customers.

Quantitative Benefit

300% five-year migration ROI from OpEx savings and cost avoidance
30% of its data was rationalized by identifying and not migrating underutilized and duplicate data
50% of the on-prem equipment was rationalized during the course of the migration