I want to implement a standard-redundant Ignition SCADA system that will read data from 337 solar inverters, distributed across twelve Sungrow data loggers. From each inverter, approximately 90 analog points will be read, using Modbus TCP/IP every 3 seconds.
The plan is to historize the data using the Ignition Historian module, writing to a MySQL database. At least 10 points per inverter will be historized every 5 seconds, while the remaining points will be historized on-change or using deadband.
The configuration I normally use is a single server where all Ignition modules are installed, along with the database. In this case, the system will be Ignition redundant. The clients (at least 2) will run on separate PCs.
My question is whether that amount of Modbus driver polling and database access from the historian can be handled by a standard Ignition installation.
Server characteristics:
HPE ProLiant DL380 Gen11
Intel Xeon Silver 4514Y — 16 cores, 2.0 GHz
64 GB RAM
RAID 5 with approximately 1.8 TB storage
Dual power supply
Same hardware for both servers in the redundant pair
I believe a large portion of your success will be based on how many requests (per device) are needed to retrieve the data you require, and how much (er, little) latency exists between the server(s) and devices.
From my experience, I’ve little concern with this amount of database throughput.
Are you able to deploy a test server on the solar network? Build a device, build and instantiate a UDT, then monitor device status to determine if the driver will sufficiently optimize requests, and if the network will support your goal.
After you have one device & tag configured, script the generation of all others…and try it?
Excellent tool, I'll definitely use it. Thanks for the info.
Another question: is it possible to manage the historical database on a NAS, for example, where we can use MySQL?