Tech Talks: Dynamic Cube Services
Join Jack Lacava, Tom Linton, and Matt Kerslake as they explore Dynamic Cube Services, which enable cube building with data and metadata from internal or external sources with data consumption via XFGetCells and Cube Views!
Key Topics Covered:
- What is Dynamic Cube Services (DCS) — building cubes with data and metadata from any source (internal cubes, custom tables, external databases, files, web services) without importing to the OneStream database
- Why use DCS — performance gains through controlled data unit sizing, reduced data duplication, greater reporting flexibility, and simpler setup for planning/forecasting scenarios
- Three core DCS components: Dynamic Dimension Service, Dynamic Data Binding (no-code), and Dynamic Data Service (code-based)
- Data Binding — no-code approach using Share Data and Copy Input Data options to create a read-only or read-write window into an existing source cube
- Member filters and pre-aggregated members — filtering by account, flow, UDs, and entities to reduce data unit size; squashing deep hierarchies into single members for faster reporting
- Dynamic Dimensions — generating account-type dimensions (Account, Flow, UD1–UD8) in code from SQL tables or external systems; auto-syncing without manual metadata management
- Service Factory architecture — how the dispatcher pattern connects workspace assemblies to Dynamic Dimension and Dynamic Data Services; shareable workspace requirements
- Setting up a Dynamic Cube — creating workspace assemblies, registering the service factory, and configuring cube properties to point to the service
- Data writeback — writing entered data back to custom tables, external databases, or external systems (not recommended for regular OneStream cubes)
- Consolidation behavior — parent-level records generated and stored in OneStream tables when using the "with calculate" data binding type
- Caching and refresh logic — configuring check intervals and change-detection logic to control dimension and data refresh cycles
- Real-world use cases demonstrated: People Planning (exposing custom table data as a cube), BI Blend integration (granular operational data exposed for cube view consumption)
- Limitations — same data unit record constraints as regular cubes; dynamic cubes are back-end/admin tools, not end-user self-service
- Expanded API documentation in v9.0 and upcoming Genesis block to simplify DCS setup in future releases