We at Hexagon/Intergraph have implemented an interesting process using FME Server and FME Engines to convert large quantities of data across multiple servers.
We are in the process of repeating this for two other major projects.
The implementation contains
1. FME Server Core (with Failover support)
2. Multiple FME Engines on 7 servers
3. C++ Front End application for job submission built with FME Server API
4. Data staged across the servers
5. Load balancing the servers
6. Check pointing the process
7. Redundancy built in
Various other parts built and automated to complete data processing consistently and accurately.
The initial project was successful without any errors over 6 months where there was new data delivered every other week.
FME World Tour 2018