Skip to main content

Connector Types

DFS Lite connectors bring source data into FactVerse workflows. Use this reference to prepare source information before creating a connector.

Connector type summary​

ConnectorTypical sourcePrepare
OPC UAIndustrial equipment, PLC gateway, historian gatewayEndpoint, security policy, namespace, node paths, credentials.
BACnetBuilding management systemDevice address, network scope, object list, polling interval.
ModbusMeter, controller, gatewayHost, port, unit ID, register map, data type, byte order.
MQTTIoT broker, edge gatewayBroker URL, topic pattern, payload format, authentication.
RESTEnterprise API, cloud system, custom serviceBase URL, endpoint, authentication, parameters, response field paths.
CSVFile import, staged export, historian extractFile source, delimiter, header rule, timestamp column, identity columns.
JDBCDatabase table or queryDatabase URL, credentials, table/query, key column, timestamp column.
FabricLakehouse or workspace dataWorkspace context, dataset/table reference, access method, owner.

OPC UA​

Use OPC UA when the source exposes industrial node hierarchy.

Required preparation:

  • endpoint URL;
  • security mode and policy;
  • certificate requirements;
  • username or credential method;
  • namespace information;
  • node IDs or browse path;
  • expected polling interval.

Validation:

  • connection test succeeds;
  • browse shows expected nodes;
  • preview returns values with timestamps;
  • source owner confirms units and data types.

BACnet​

Use BACnet for building automation and BMS point data.

Required preparation:

  • BACnet/IP route or gateway;
  • device ID or discovery scope;
  • object type and instance list;
  • polling interval;
  • unit expectations;
  • network access approval.

Validation:

  • devices are reachable;
  • objects resolve to expected names;
  • values update at expected cadence;
  • polling stays within building network expectations.

Modbus​

Use Modbus for meters, controllers, and gateway data with known register maps.

Required preparation:

  • host;
  • port;
  • unit ID;
  • register address;
  • function code;
  • data type;
  • byte order;
  • scale factor;
  • polling interval.

Validation:

  • sample register values match source-owner reference values;
  • signedness and byte order are correct;
  • transformed values are within expected range.

MQTT​

Use MQTT for topic-based IoT and edge data.

Required preparation:

  • broker URL;
  • port;
  • TLS requirements;
  • credential or certificate;
  • topic pattern;
  • payload format;
  • timestamp field;
  • quality of service expectation.

Validation:

  • subscription receives messages;
  • payload parses correctly;
  • identity, timestamp, and metric fields are present;
  • topic naming is stable.

REST​

Use REST when source data is available through an API.

Required preparation:

  • base URL;
  • endpoint path;
  • authentication method;
  • query parameters;
  • pagination behavior;
  • response field paths;
  • timestamp field;
  • rate limit;
  • expected request cadence.

Validation:

  • test request succeeds;
  • response shape matches mapping needs;
  • pagination preserves records;
  • error responses are understandable to operators.

CSV​

Use CSV for first imports, offline exports, and staged integration.

Required preparation:

  • file source or upload process;
  • delimiter;
  • header row rule;
  • encoding;
  • timestamp column;
  • identity column;
  • unit columns;
  • expected refresh or replacement behavior.

Validation:

  • preview shows correct columns;
  • special characters parse correctly;
  • timestamp and numeric values parse correctly;
  • row count matches the source export.

JDBC​

Use JDBC when source data is available from a database.

Required preparation:

  • database URL;
  • driver and connection method used by the deployment;
  • credential;
  • table or query;
  • primary key;
  • timestamp column;
  • incremental sync condition;
  • read-only account approval.

Validation:

  • query returns expected rows;
  • incremental condition is correct;
  • source indexes support the query;
  • database owner approves the access pattern.

Fabric​

Use Fabric when project data is prepared in a lakehouse or workspace context.

Required preparation:

  • workspace or lakehouse reference;
  • dataset or table name;
  • access identity;
  • owner;
  • refresh cadence;
  • schema owner;
  • expected downstream workflow.

Validation:

  • dataset can be listed;
  • preview returns expected rows;
  • schema owner confirms table meaning;
  • refresh cadence matches the DFS workflow.

Naming convention​

Use connector names that identify source, site, and purpose.

Examples:

  • SG facility BMS BACnet
  • Plant historian OPC UA
  • CMMS work orders REST
  • Energy meter CSV import

Next step​

Use Connector Configuration for detailed setup steps.