System Requirements
| Recommendation |
---|---|
CPU | Modern x86-64 CPU (Intel Xeon, Intel Core, AMD Ryzen) |
Operating System | Windows, Linux, Mac |
CPU Power | 2GHz+ |
RAM | 4GB (8-16GB+ Recommended) |
Flash | 32GB SSD |
Interface | Gigabit LAN |
Project Sizing:
RAM (Most IMPORTANT for Sizing):
The most important resource for server sizing is RAM, which must properly tune your JVM heap size for any medium to a large project.
A good rule of thumb is to plan for 1GB to 2GB of heap for every 10K points.
Heap size of more than 1.5GB will need to run a 64-bit JVM (the actual max heap size of a 32-bit JVM will vary slightly).
85% of Java Heap consumption is roughly our recommended max before considering upgrading the amount of space.
Open JDK is recommended, but you can also use Java 11
Another important aspect to consider is that the historian is limited mainly by I/O performance (disk drive speed). Using an SSD will likely have a significant performance boost for reading and writing history data. Due to Folio's highly efficient compression techniques, disk space is typically not an issue for projects. Analysis of projects has shown an average of about 1.5 bytes per timestamp sample, equating year per point for 15min interval data.
Depending on the apps/features you use (and don't use) will significantly impact performance and how much horsepower you will need to run smoothly.
curVal synchronization and his collection typically requires a high amount of bandwidth and CPU processing
batch synchronization of history data is typically far more efficient and easier to tune than curVal synchronization
how many programs/alarms is will have a significant impact on CPU processing
running multiple projects on a single host consumes more overhead; 100 little projects of 100 points require more horsepower than a single 10K point project
Max amount of points that should be in the watch at one time should be around 50k
How points get into a watch:
Trend (Always)
Programs (Always)
Alarms (Always)
Graphics (On View Only)
Point Graphics (On View Only)
Summaries (On View Only)
Database limits
Maximum 300 connectors per FIN instance
Maximum 2000 items per tree