When it comes to collecting data not all systems are created equal. Systems with open services make it far easier to extract and transmit their data. Here are some notes on systems that we prefer, because they include Servers or services with self-describing schema:
• BACnet – device and object descriptions within the protocol (down to units and functional descriptions).
• OPC – device and object descriptions. Simple shallow layers. OPC XML DA is particularly attractive.
• Sunspec – layers on REST or Modbus, with a standard, globally agreed upon schema.
• RESTful – MTConnect – globally agreed upon schema, made simple by uni-directional function.
We can also use servers or services with some sort of external schema (These above-mentioned services can also be converted to a service with self-describing schema):
• Modbus (requires the register map for any device).
• SNMP (requires the resource map).
• Automation control system protocols – like Ethernet/IP and various proprietary protocols.
We can also use message queues:
• and so on.
These too need a schema, but have an “always on and always flowing” character that makes dabbling easy. With the right keys, these are sometimes even convertible/usable by looking at the context of the flow itself.
What about SQL? It can sometimes be an open or authenticated service. Database formats need their schema to be able to be communicated via some standard (or simple) process. There is no general system for such schema exchange.
The worst scenario is when DB schema tend to be optimized for internal manipulation but NOT data exchange.
Comments will be approved before showing up.