Skip to main content

Four Open Standards for Data Management

Which open standards describe your data, and how do they fit together?

By Dr. Simon Harrer · · 1:49 video · 7 slides

In short

Four open standards describe your data on three levels. ODPS describes the data product, and ODCS describes its data contracts. Apache Ossie describes semantic models and the ontology. Each column in a data contract points to its business concept.

Watch on YouTube
Slide 1 of Entropy Explains episode 13, Files that every tool reads: Define once, use everywhere only works if all tools read the same files.

Files that every tool reads

Define once, use everywhere only works if all tools read the same files. There is one file for the data product, one for the data contract, one for the semantic model, and one for the ontology. BI tools, data platforms, catalogs, and AI agents read these files.

Slide 2 of Entropy Explains episode 13, Three levels: The standards work on three levels.

Three levels

The standards work on three levels. The physical level has the tables. The logical level has data products, data contracts, and semantic models. The conceptual level has the ontology, which says what your business means.

Slide 3 of Entropy Explains episode 13, ODPS: the data product: ODPS, the Open Data Product Standard, describes the data product.

ODPS: the data product

ODPS, the Open Data Product Standard, describes the data product. It names the owner, for example the checkout team. It also lists the input ports and the output ports. Consumers and other data products connect through these ports.

Slide 4 of Entropy Explains episode 13, ODCS: the data contract: ODCS, the Open Data Contract Standard, specifies one output port.

ODCS: the data contract

ODCS, the Open Data Contract Standard, specifies one output port. The data contract contains the schema, the quality rules, the service levels, and the terms of use. Bitol, a Linux Foundation project, maintains ODCS and ODPS.

Slide 5 of Entropy Explains episode 13, Ossie: the semantic model: Apache Ossie adds two layers that carry meaning.

Ossie: the semantic model

Apache Ossie adds two layers that carry meaning. The semantic model is on top of the tables and is for analytics. It contains datasets, joins, and metrics. Every tool then calculates revenue in the same way.

Slide 6 of Entropy Explains episode 13, Ossie: the ontology: The ontology describes the business itself, without tables.

Ossie: the ontology

The ontology describes the business itself, without tables. It contains concepts, relationships, and rules. A customer places orders, and an order total is never negative. This is true on every platform where the data is.

Slide 7 of Entropy Explains episode 13, How they fit together: The ontology defines the meaning.

How they fit together

The ontology defines the meaning. Data products and semantic models implement it. The data contract connects the two levels: each column points to the concept that it implements. Here, the column customer_id points to the concept Customer.

Read more