A nonprofit organization responsible for the distribution of digital performance royalties selected Neovera to deliver a solution that would help optimize their custom application. The application manages the collection of digital performance records from its clients as well as the distribution of royalty payments to the necessary artists. This process of distributing royalties requires a series of steps to merge data feeds and link that data with artists and labels to determine proper royalty payments each quarter. This was taking a better part of the quarter to process, which resulted in little time between distributions. It was also anticipated that larger data demands could limit the amount of distributions that could occur in this timeframe.
The Neovera team conducted a thorough analysis of the organization’s infrastructure, including an assessment of the existing Linux Red Hat servers using a targeted tuning approach. We used previous run times for the application to measure and target the most time consuming processes and to focus the tuning approach on those elements that would produce the greatest impact. Neovera performed a comprehensive tuning analysis for the following:
- Operating System Tuning – I/O wait times, Qdepth size, CPU wait times
- Storage Tuning – ASM LUN layout
- Network Tuning – Kernel parameters for public and private network
- Database Tuning – Database parameters, parallel processing, partitioning
- Application/SQL Tuning – Analyzing index usage and eliminating unused or redundant indexes and adding indexes to match application usage.
Additionally, Neovera recommended a transition to Oracle RAC, which would enable the company to cluster Oracle databases and bind multiple servers so they operate as a single system. We then architected and implemented the switch to the new platform.
BOTTOM LINE RESULTS – OVER 70% REDUCTION IN PROCESSING TIME
Neovera’s solution exceeded performance requirements with over a 70% reduction of total application processing time, thus allowing more data to be processed in less than one day, and facilitating the distribution to occur monthly instead of quarterly. In fact, the processing time for one of the larger activities dropped from 50 to 11 hours. A smaller procedure was similarly reduced from 6 hours to just 3 minutes.