Metadata Provenance

The open metadata ecosystem draws together metadata from many sources. Metadata provenance provides information about where metadata has come from and how it can be maintained (that is updated and deleted).

Metadata Collections

The metadata that a specific technology instance creates and maintains is collectively referred to as a metadata collection. Each metadata collection has a unique identifier, called the metadata collection id and an optional metadata collection name. When metadata from the collection is shared with the broader open metadata ecosystem, it includes the metadata collection id and metadata collection name in its header along with a category name that describes how the metadata was introduced into the open metadata ecosystem. This category name is called the Instance Provenance Type by the Open Metadata Repository Services (OMRS) and Element Origin by the Open Connector Framework (OCF) and most of the Open Metadata Access Services (OMASs). We will use “Element Origin” for the rest of this description since it is the most commonly used name.

Figure 1 shows the different values of Element Origin and how they tie into the mechanism used to introduce the metadata to the open metadata ecosystem.

Figure 1

Figure 1: What the different values of Element Origin say about the source of metadata

Using provenance to manage metadata integrity

Egeria ensures that only the owner of a metadata instance is permitted to update it. This enforcement makes use of the metadata provenance information in the metadata instance’s header.

Typically the owner is the originator of the metadata instance, but the section below describes how to move a metadata instance’s home from one metadata collection. This effectively changes the owner to the new metadata collection.

Changing the metadata collection that a metadata instance belongs to

The Open Metadata Repository Services supports commands to change the metadata collection that a metadata instance belongs to. This should be done only if the instance needs to be edited and the technology supporting the original metadata collection is no longer available. For example, for instances belonging to a deregistered repository. The change needs to be made with care and planning ensuring that all members of the cohort are connected when the command is issued so that the change of ownership can be recorded consistently in all repositories.

Further Information

License: CC BY 4.0, Copyright Contributors to the ODPi Egeria project.