We are now converting the National Korean Claim data into CDM. The National Health Insurance Service, the only national compulsory insurer in Korea, has the data. We need to deploy the converted data into a very fast DBMS. I recommended them Microsoft PDW, because I know that CDM works perfectly with the PDW. However they already have Oracle Exadata and SAP HANA and want to use one of them for CDM DB.
Does anyone have successfully used these DBs for CDM?
Hi Rae,
We have not used Oracle Exadata for hosting OMOP CDM data but in my past role/company we used Oracle Exadata to host some pretty big data sets. It is a scalable platform, yet pretty pricey, as you would expect from a proprietary appliance.
Sizing an environment like this would of depend on a very specific type of Exadata you have, rack configuration and system capacity e.g. number of CPU/cores, RAM, raw/available disk space, nodes in the rack etc… When I first was exposed to Exadata they were in X2 going into X3. It seems that they are on X6 version now. What is their Exadata configuration?
You would then need to match the size of your data set (including any additional growth, future archiving/backup and analytical needs) with system capacity you have. How big is your target data set (in patients or Gb/Tb)? I would assume that the National Korean Claims data would be quite big.
I got to admit I am not up to date on the latest Exadata developments, but one of the challenges with appliances like that was growing storage capacity when we needed it.
Hi Greg,
Thank you for the information and consideration points
I will discuss with the engineer of the NHIS with the points, and will try to deploy a small pilot project with it.
Thank you.
Rae