The electronic component manufacturer moves from Azure to Oracle Cloud Infrastructure and Exadata Cloud Service, saving 17 hours in daily reporting.
“With Oracle Cloud infrastructure and Exadata Cloud Service, we were able to reduce the time for daily data warehouse reporting by 75%. Now we can meet our business users’ needs for accurate inventory data.”
KEMET produces an extensive portfolio of electronic components that are used in everything from spacecraft to defibrillators to products we use every day. It ships more than 50 billion components per year to 180,000 customers around the world. To support this high-volume business, it uses Oracle E-Business Suite applications supported by a data warehouse on Oracle Database.
Recently KEMET migrated that data warehouse from its own data center to Azure cloud service, but that had led to a deterioration of performance. Its business users were no longer able to get an up-to-date, accurate view of their inventory data, and that impacted their business efficiency.
KEMET was committed to its decision to retire its own data center, but needed a better cloud solution than Azure could provide.
Why KEMET Chose Oracle
KEMET learned from the experience with Azure that it is critical for the performance of its data warehouse that it has access to dedicated compute and storage infrastructure. Given that the sources for its data uploads are distributed geographically, it was also important that Kemet could optimize the network connectivity between these locations.
Oracle Cloud offers bare metal compute instances with attached NVMe solid state drives (SSD) that provide the performance and control that Kemet needs. Using Oracle Cloud FastConnect, Kemet could also implement an Oracle SD-WAN that satisfied its connectivity needs. Using Oracle Exadata Cloud Service consolidated transactional and data warehouse workloads for greater efficiency and performance and simplified support.
Once KEMET migrated its data warehouse from Azure to Oracle Cloud Infrastructure and Oracle Exadata Cloud Service for increased performance, it realized 50% lower CPU consumption compared to the previous commodity servers on-premises. By implementing Oracle SD-WAN optimizations, the company was able to get a drastic improvement in performance (in batch job runs). Kemet’s inventory load job, which took about 5.3 hours to run in Azure, completed in less than an hour in OCI.
Another job, which took about 12.5 hours to run in Azure, completed in less than 2.15 hours in OCI. Kemet’s overall BI batch job schedule run, which ran in 23.5 hours in Azure, now completed in only about 6 hours.
This performance was also better than what KEMET was getting from its own data center before the migration to Azure.
KEMET’s IT teams are now able to meet their SLAs.