More than a year ago, as part of building our Clinically Integrated Network, we ran into performance issues when working with 5 billion records containing de-identified healthcare data. Loading the data to a SQL Server database in Azure took 7 days, and some large queries afterwards were running for 9 days. Having seen quite a few articles on Snowflake at that time, we decided to explore it. There was no need to wait for weeks to go through multiple levels of approval. Once set up, loading the data to a Snowflake database took just 1 hour. By simply tweaking the warehouse size, our large queries ran in 30 min. Incredibly fast. At the same time, it only cost us about 1/10^th of what it did using SQL Server and did not require any additional resources to optimize the system. With these facts, I was able to easily make a case to our Executive team that we needed to use Snowflake for processing massive data sets. A quick discussion was all that was required for us to proceed.

Although not a comprehensive comparison between the two database systems and really specific to our use case, it served us very well to evaluate a new technology like Snowflake. So, while we will be using SQL Server on Azure for most of our core projects, we expect to continue leveraging the power of Snowflake for analyzing massive datasets that will allow our power/business users to do their analysis without needing to wait too long. Without the latitude and the ownership available to explore and make such decisions at MPOWERHealth, we may not have been able to take advantage of some of these new technologies that greatly benefit us and our customers.