Novant Health
Case Study
Faster, easier data access and analytics
The leading online ordering and delivery giant with a growing cloud data lake switched from EMR to Starburst Enterprise to improve overall performance and efficiency.
100%
of data in Amazon S3
10,000
daily inquiries
250 TB
data processed daily
Region
Americas
Industry
transportation
Environment
aws
Solution
enterprise
Employees
1000+
Anonymous
Data Engineering Manager
Fortune 500 Oil and Gas Company
“We primarily use Starburst for reporting and dashboard queries, but the platform supports a variety of use cases, including ad-hoc queries, complex queries, and some exploratory, machine learning, and data science activity. We’ve even started using it to serve A/B testing experiment results and for real-time reporting and visualization.”
About
The country’s leading online and mobile food ordering and delivery marketplace operates in over 2,200 U.S. cities and London. The company processes 521,000 daily orders and serves 19.3 million active diners.
This activity drives tremendous data growth. To remain the leader in an increasingly competitive market, especially after the pandemic-fueled uptick in mobile ordering, the company needed to make its data available immediately for internal analysis.
Challenge
Not long ago, the company had a major data center footprint and relied on an expensive Java and MySQL stack. The company streamed data into an AWS data lake, but open source Trino on Amazon EMR failed to deliver the necessary performance and efficiency. Unsuccessful queries that occurred when using EMR were driving up costs.
As the business rapidly grew, the company recognized that this approach would not scale cost-effectively. It needed enterprise-grade software and a more available, scalable, and flexible system.
Solution
The data engineering team selected Starburst Enterprise as the query engine to help users transform raw data into actionable business intelligence. The company operates a cloud-only environment with 100% of its data in its Amazon Simple Storage Service (S3) data lake, and uses Starburst for reporting and dashboard queries, with 90 to 95% of analytics coming from Starburst.
Along with two production clusters in AWS, the data engineering team uses on-demand Starburst Enterprise clusters for job-specific workloads throughout the day. This allows the team to dynamically spin up a cluster, run queries, and then power it down again to minimize compute costs.
Results
Deploying Starburst Enterprise resulted in these direct benefits to the company.
- Faster performance and a higher successful query rate, leading to additional adoption by end users.
- Reduced customer cost due to fewer unsuccessful queries.
- Cluster optimization is based on use cases, allowing each cluster to perform better.
- Improved security and stronger controls around data access that is essential for publicly traded companies.
As the company works to secure its leadership in an increasingly competitive market, Starburst Enterprise provides the data engineering team with the reliable, scalable, high-performing, cost-effective platform it needs to support end users and extract business intelligence.
