Most organisations don’t get to start with a blank slate. A full migration from a legacy data platform to Fabric is a significant commitment, and often the appetite simply isn’t there yet. So how do you start getting value from Fabric today, without waiting for a migration that may never come?
Fabric can deliver on the following without migrating your existing platform:
If your organisation has already purchased an F64 Fabric capacity to distribute Power BI reports without requiring individual Pro licences, there is a reasonable chance that unused compute is sitting idle. Check the Fabric Capacity Metrics app to assess headroom before purchasing anything additional.
If no Fabric capacity exists yet, it does not need to be expensive. For pipelines where Fabric is orchestrating jobs on other database technologies rather than doing heavy compute itself, a small capacity such as an F2 is often sufficient. Fabric capacity can be paused when not in use, further minimising costs.
If your legacy platform runs batch loads against a SQL database, those loads can be orchestrated using Fabric pipelines. Building metadata-driven orchestration patterns is significantly more straightforward in Fabric than in older tooling.
When a new data source needs to be integrated into your platform, Fabric is a natural starting point. Fabric pipelines can extract data from the source system and land it into a Fabric Lakehouse, from which it can be pushed into the dimensional model on the legacy platform, or combined with legacy data directly inside a Power BI semantic model.
For teams that prefer to keep all data on the legacy platform, pipelines can also push data straight into the legacy staging layer and then trigger the downstream load from there.
One area where Fabric holds a clear advantage over legacy tooling is streaming data. Many streaming use cases are isolated from batch-loaded dimensional models, making them an ideal candidate for an initial Fabric workload. Streaming data landed in Fabric using Eventstream can be enriched with batch data from the legacy platform inside a Power BI semantic model, delivering a near-real-time view without disrupting existing architecture.
Fabric pipelines can connect to on-premises or IaaS hosted platforms securely using the Microsoft on-premises data gateway. A cluster of gateways should be configured for high availability.
For cross-platform dependency management and unified error notification, it is useful to bring Fabric and legacy workloads into the same scheduling framework. This can be achieved in either direction.
Fabric pipelines can be triggered from a legacy orchestration engine by calling the Fabric REST API. The API returns pipeline run status and error detail, which can be read by the legacy scheduler for consolidated logging and alerting.
Alternatively, Fabric pipelines can orchestrate legacy workloads directly, handling complex dependency chains that span both platforms within a single pipeline.
Some upfront design work on your Lakehouse structure will pay dividends as your Fabric footprint grows. Whether sources from different domains should share a single Lakehouse or be separated into distinct ones depends on whether the domains are related, security requirements, team ownership, and how the data will be consumed downstream. File structure conventions within the Lakehouse should be agreed on before data starts landing.
Fabric can integrate with the same Git providers as your legacy data platform, though it is advisable to create a dedicated repository for Fabric assets. Deployment can be handled through existing CI/CD pipelines, or through Fabric’s native deployment pipelines for promoting content across development, test, and production environments.
Fabric works well alongside legacy platforms, making it a natural starting point for modernisation rather than an all-or-nothing commitment. Pipelines can sit alongside your legacy workloads, new data sources can land in a Lakehouse without touching your existing platform, and orchestration can span both platforms. With some upfront design work, this becomes a strong foundation for additional workloads or, when the time is right, a full migration.
If your organisation is looking to get started with Fabric alongside a legacy platform, get in touch with the One51 team.