Supported Platforms
Cloud
ModelOps is only supported on cloud infrastructure.
This release of ModelOps is verified on these cloud environments
The Kubernetes clusters must be running Kubernetes version 1.21 or later with Linux nodes.
There is limited support for mixed Linux and Windows clusters for the Statistica model runner. In this case the Statistica Scoring Server node is supported on Windows (Azure only, no support for Windows nodes on EKS or OpenShift). No other ModelOps servers are supported on Windows.
Authentication Providers
ModelOps supports these authentication providers:
Data Channels
ModelOps data channels support:
- Apache Kafka version 2 or later (Kafka Data Channels)
- any JDBC compliant database (JDBC Data Channels)
Models
ModelOps supports these model types for scoring:
- PMML - Version 3 and 4
- Python (both binary and script) - Version 3.8
- Spark ML - Version 2 and 3
- Statistica1 - Version 14
- TensorFlow2 - TensorFlow 2
Integrations
- TIBCO ModelOps Integration with TIBCO Spotfire requires TIBCO Spotfire 11.4.0 LTS or higher
- TIBCO ModelOps Integration with JupyterLab requires JupyterLab 3.3.0 or higher
Notes
1 Statistica workspaces having: the input and output nodes marked; and not exercising Statistica Enterprise functionality.
2 GraphDef and SavedModel formats stored as binary Protocol Buffers.