Data stores (early access)

This release of CONNECT is available to selected customers. For more information about this release and its availability, please contact your sales representative.

A data store is a cloud-based storage solution built from specialized storage service instances, each handling a different kind of data object. It provides one place to manage the operational data used by industrial applications. Some related services are also provisioned along with the data store to move, enrich, or publish data, but those services are related to data connections or data modeling, not storage.

Supported resources

This data store supports the following resources:

• Streams: Sequential and time-series records used to capture ordered data over time, such as sensor readings.

  Note: Streams use their own type model and does not depend on the Types mentioned below. Event Management and Entity Management, by contrast, depend on Types for schema validation.

• Entities: Long-lived business resources and relationships that represent operational context.

• Types: Reusable type definitions that provide structure and validation rules for typed data, used by events and entities.

• Events: Time-bounded operational records that capture when something happened and its lifecycle state.

These different resources are stored in independent service instances for storage, but CONNECT presents them as one seamlessly managed data store. This service separation supports different access patterns, lifecycle rules, and scaling needs for each data category, which helps prevent one type of data activity from degrading another and allows each service to evolve independently.

Partioning and regional isolation

Storage is also partitioned by account and geography to enforce isolation boundaries, improve operational safety, and enable targeted scaling and performance tuning with clear ownership and data residency controls. In practice, each geography has one shared set of core service instances used by all sites in that geography.

Data store provisioning

When you create a site, CONNECT determines whether to create a regional data store based on geography.

For each account and geography, CONNECT uses one shared regional data store. The first site in that geography creates it. Later sites in that same geography reuse it.

For example, a school district might have separate sites for schools, campuses, and administrative offices in one region. Those sites remain separate, but they share the same regional data store for that geography.

Provisioning the first site in a new geography

When you provision the first site in a geography, CONNECT provisions the regional data store for that geography. This includes shared core storage services (streams, types, entities, and events), related supporting services, and security context for authorization and data ingress.

Provisioning an additional site in an existing geography

When you provision another site in that same geography, CONNECT provisions the site and its site-level security context, then links it to the existing regional data store.

No new data store is created in this case. The site uses the existing shared regional data store that was created for the first site in that geography.

This model keeps onboarding predictable, avoids duplicate regional infrastructure, and provides a consistent data store boundary across sites in the same geography.

Data ingress and egress

For a data store to deliver value, data must first be written into it and then egressed for use in analytics and other downstream applications. In other words, ingestion captures operational signals, and egress turns those signals into business outcomes.

Ingress is the process of bringing data into the data store from external systems, devices, and applications so it can be stored and managed in CONNECT. For more information about data store ingress and ingress options, see Data connections (early access).

Egress is the process of reading, exporting, or publishing data from the data store so it can be consumed by analytics and other value-add applications. For more information on CONNECT features related to use of egressed data, see Data modeling (partner preview).