Open Metadata Repository Cohort

An open metadata repository cohort is a collection of servers sharing metadata using the Open Metadata Repository Services (OMRS). This sharing is peer-to-peer. Once a server becomes a member of the cohort, it can share metadata with, and receive metadata from, any other member.

OMRS needs to be flexible to support different performance and availability requirements. For example, where metadata is changing rapidly (such as in a data lake), this metadata should be dynamically queried from the repository where it was created and is being maintained because the rate of updates mean it would cost a lot of network traffic to keep a copy of this metadata up to date. The repository where a piece of metadata (metadata instance) was created and where it is maintained is called its home metadata repository.

On the other hand, governance definitions (such as policies) and glossary terms rarely change. They are often administered centrally by the governance team and then linked to all metadata that describes the organization’s data resources. Thus it makes sense for this metadata to be replicated across the repositories within the cohort. These copies are called reference copies of the metadata and they are read-only.

The role of the OMRS is to optimize access to metadata across the cohort by using a combination of replication and federated queries, driven by the the metadata workload from the connected tools.

Cohort membership

To join an open metadata repository cohort, a metadata repository must integrate with the OMRS module. Egeria provides a number or pre-built cohort members.

One of them, the repository proxy provides a simple way to integrate a third party server into a cohort by creating an OMRS Repository Connector and optional Event Mapper Connector to map between the third party APIs/events and the repository service’s equivalents

A more bespoke integration involves:

There are different integration patterns available to help you choose the best approach for your product. Each method is optimized for specific use cases and so the metadata repository can only play a full role in the open metadata use cases if it supports all integration methods. These are:

Cohort registration

The OMRS protocols are peer-to-peer. Each repository in the cohort has an OMRS Cohort Registry that supports the registration of the metadata repositories across the cohort. Through the registration process, each OMRS Cohort Registry assembles a list of all of the members of the cohort. This is saved in the OMRS Cohort Registry Store. The list of connections to the remote members of the cohort are passed to the OMRS Enterprise Connector Manager that in turn manages the configuration of the Enterprise OMRS Repository Connectors.

The Enterprise OMRS Connector provides federated query support across the metadata cohort for the Open Metadata Access Services (OMAS).

When a metadata repository registers with the OMRS Cohort Registry, the administrator may either supply a unique server identifier, or ask the OMRS to generate one. This server identifier (called the metadata collection identifier) is used in the OMRS event notifications, and on OMRS repository connector calls to identify the location of the home copy of the metadata entities and to identify which repository is requesting a service or supports a particular function.

Once the metadata repository has registered with the OMRS Cohort Registry, it is a member of the metadata repository cohort and it can synchronize and share metadata with other repositories in the cohort through the OMRS Topic.

Note: A single metadata repository can register with multiple metadata cohorts as long as its server identifier is unique across all cohorts that it joins and it manages the posting of events to the appropriate OMRS Topic for each cohort it register with. This capability is useful for a metadata repository that is aggregating reference copies of metadata from multiple open metadata repository cohorts.

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