Case Studies > Consolidating Many Systems to One Modern Data Lake for Devon Energy

Consolidating Many Systems to One Modern Data Lake for Devon Energy

Customer Company Size

Large Corporate

Region

America

Country

United States

Product

Snowflake
Devon Data Hub
Attunity
Databricks
Power BI

Tech Stack

Cloud Data Platform
Data Lake
Data Warehouse
ODBC
JDBC

Implementation Scale

Enterprise-wide Deployment

Impact Metrics

Productivity Improvements
Cost Savings
Digital Expertise

Technology Category

Infrastructure as a Service (IaaS) - Cloud Databases
Platform as a Service (PaaS) - Data Management Platforms
Analytics & Modeling - Big Data Analytics

Applicable Industries

Oil & Gas

Applicable Functions

Business Operation

Services

Cloud Planning, Design & Implementation Services
System Integration

About The Customer

Devon Energy is a prominent independent oil and natural gas exploration and production company based in Oklahoma City, Oklahoma. The company is dedicated to meeting the world's growing energy demands through its onshore operations in the United States. Devon Energy's portfolio includes a stable and environmentally responsible production of oil and gas properties, providing a platform for future growth. The company produces approximately 140,000 barrels of oil per day, along with 600 million cubic feet of natural gas and 75,000 barrels of natural gas liquids daily. With a deep inventory of development opportunities, Devon Energy is well-positioned to deliver future growth in the energy sector.

The Challenge

Prior to implementing Snowflake, Devon Energy faced significant challenges with its legacy systems. The company attempted to build an enterprise data warehouse three times, each time encountering issues that hindered scalability and efficiency. The first attempt failed due to the system's inability to scale with high data and query volumes, reaching its capacity limit within six months. The second attempt required full-time development staff for maintenance and had numerous points of failure, leading to increased costs. The third attempt lacked visibility into user activities, resulting in duplicated queries and inefficiencies.

The Solution

To address the challenges faced with its legacy systems, Devon Energy selected Snowflake as the foundation for its new enterprise data warehouse, known as the Devon Data Hub. This modern data hub integrates a data lake and data warehouse, capable of handling structured, semi-structured, and unstructured data. The Devon Data Hub ingests raw data from over 30 systems, encompassing more than 50,000 tables and 40+ TB of data, supporting over 1,000 users and processing 4 million queries per month. Snowflake's architecture allows for the separation of storage from compute, enabling users to access all enterprise data without IT intervention. The data hub features two layers: an enterprise data layer with robust governance managed by IT staff, and a community data layer with minimal governance to facilitate ease of access for citizen developers. This setup allows users to connect raw source-system data to curated enterprise data seamlessly.

Operational Impact

The Devon Data Hub has democratized data access across the organization, allowing all employees to query 95% of the enterprise's data.
The modern cloud architecture of the Devon Data Hub supports structured, semi-structured, and unstructured data, enhancing data handling capabilities.
The separation of storage from compute in Snowflake allows for flexible scaling to meet business demands while controlling costs.
The integration of tools like Attunity, Databricks, and Power BI with Snowflake enhances data replication, transformation, and user access.
The Devon Data Hub's two-layer structure supports different user classes, with robust governance for enterprise data and minimal barriers for community data access.

Quantitative Benefit

A regulatory report that previously took 48 hours in SQL Server now runs in just minutes.
A data preparation process that took 15 hours before now runs in 30 minutes.
2,000 simultaneous and random queries against a 40-billion-record table return results in under 10 seconds.