How We Modernized Our Data Warehouse Using Data Virtualization
At Vancouver Coastal Health, we provide health care services to roughly a million people through a broad network of hospitals, primary care clinics, community health centers, and residential care homes. I lead the Decision Support team, which delivers analytical and planning support to operational, clinical, and corporate departments within the organization. My team is responsible for large volumes of data, and for many years, we relied on our data warehouse to support our customers. When it came time for our main patient information system to be replaced (a large, multi-year project), we also decided that it was time to modernize our data architecture.
A Well-Needed Upgrade
One of the main challenges was combining the new highly transactional system with our dimensionally modeled, legacy data warehouse (implemented in SQL Server). The solution would need to enable us to report on data that spans both source systems and support rapid development, as well as to provide data lineage and meta data management. Also, we needed a way to test the new infrastructure as we were phasing it in, and our data warehouse would not support that.
After looking at a range of different modernization options, we chose the Denodo Platform, which uses data virtualization to implement our core use case, that of integrating our data warehouse with the new transactional system and providing close to real-time reporting.
A Powerful Architecture
The architecture of data virtualization was a perfect fit for our needs. Instead of replacing our legacy data warehouse, we implemented data virtualization as an organization-wide layer above both the data warehouse and the new system. Rather than replicating data from both systems into a third R/DBMS, data virtualization provides integrated, virtual views pointing to both systems, which enable customers to report on all of our data without moving the source data using traditional ETL processes. This ability to leave the data in the source until it is requested by the end user enables us to provide data closer to real time than a typical batch-oriented process. We can provide reports that are about 10 minutes out from the last transaction. The Denodo Platform is set up to automatically transform views of the new transactional system so that they mimic the structure of our data warehouse.
Security is both strengthened and simplified, as the Denodo Platform provides a single point of entry to all the data in our environment. This enables us to implement flexible, powerful rules governing which users can see or access which data, which is extremely important in any healthcare environment.
Another benefit has been automated testing. Built into the Denodo Platform, this capability enables us to run thousands of tests every night to make sure that any new code is free of errors, that the data is accurate, and that no one has made any changes under the hood without us knowing.
I sat down and told our modernization story in the video below, and I hope you will have a listen. I strongly recommend data virtualization as an efficient and powerful data warehouse modernization strategy.
To learn more about how companies are leveraging data virtualization across numerous industries, see the collection of Denodo case studies.
- How We Modernized Our Data Warehouse Using Data Virtualization - September 11, 2019