Landauer (Talend) achieves 50% infrastructure cost reduction and 80% lower latency by migration to AWS OpenSearch
About Landauer - Talend
LANDAUER is the leading global provider of technical and analytical services to determine occupational and environmental radiation exposure and the leading domestic provider of outsourced medical physics services.
The company provides dosimetry services to about 1.8 million individuals globally and is used by 78% of United States hospitals. The company offers tools and support for organisations with potential exposure to ionizing radiation, helping them achieve their radiation safety goals.
LANDAUER’s innovations in radiation safety continue to shape the industry in dosimetry and medical physics.
The Challenge
Landauer faced several daunting challenges in optimizing their ETL workflow. First and foremost, they needed to maintain the order of processing to ensure that parallel iterations could be processed at a time.
Additionally, Landauer needed to transition away from Talend and build a new workflow using AWS Step Functions. This was a significant undertaking, requiring a thorough understanding of the new system and its capabilities. The new workflow required to be able to handle the demands of Landauer’s high-volume data processing needs, all while optimizing the cost of the workflow and providing automation, speed, and reliability.
Given these complex challenges, Landauer needed a partner with deep expertise in cutting-edge serverless frameworks to help them save costs and scale better. That’s where AntStack came in!
Our Goals
The primary objective of the project was to help Landauer transition from their existing ETL process using Talend to a more cost-effective, reliable, and efficient workflow using AWS Step Functions.
The project aimed to:
Optimize the workflow cost while delivering top-notch automation, speed, and reliability
Ensure the new system can handle the demands of high-volume data processing requirements
Provide parallel processing for packages and runs to reduce the time for data processing
Automate the entire process from development to deployment while still providing the ability to alert in case of failures
Maintain the order of processing to ensure that independent data iteration could be processed at the same time
Create a more scalable and flexible architecture that would be sustainable for future growth and expansion

Comments
Post a Comment