Microsoft Fabric decision guide: Choose a data movement strategy

Microsoft Fabric offers several ways to bring data into the platform. This guide gives you a clear recommendation for the most common scenarios, then provides detailed feature comparisons when you need to dig deeper.

Quick recommendation: Which data movement option should I use?

Use the medallion architecture as your guide:

• Gold data (reporting and analytics on processed data) - Use Mirroring. If you already have ETL processing elsewhere and mainly need to bring curated data into Fabric for reporting, Mirroring is the simplest and most cost-effective choice. It's free, requires minimal setup, and continuously replicates your data into OneLake.

• Bronze data (raw ingestion) - Start with Copy job. When you're ingesting raw data, you'll quickly need transformations, schema mapping, scheduling control, and incremental loading. Copy job gives you these capabilities natively without the complexity of building pipelines.

• Real-time streaming data - Use Eventstreams. For low-latency, event-driven ingestion and processing, Eventstreams provides real-time pipelines with no-code transformations and routing to multiple destinations.

• Complex orchestration - Pipelines will give you the orchestration flexibility you need, and copy activities in pipelines offer data object parameterization, and metadata driven data ingestion. Otherwise, copy job activity and copy activity are equitable in a pipeline.

For a full side-by-side breakdown of capabilities and supported features, see the detailed feature comparison.

Key concepts

• Mirroring gives you a simple and free way to mirror operational data into Fabric for analytics. It's optimized for ease of use with minimal setup, and it writes to a single, read-only destination in OneLake.

• Copy activities in Pipelines is built for users who need orchestrated, pipeline-based data ingestion workflows. You can customize it extensively and add transformation logic, but you need to define and manage pipeline components yourself, including tracking the state of the last run for incremental copy.

• Copy Job makes data ingestion easier with native support for multiple delivery styles, including bulk copy, incremental copy, and change data capture (CDC) replication, and you don't need to build pipelines, while still giving you access to many advanced options. It supports many sources and destinations, and works well when you want more control than Mirroring but less complexity than managing pipelines with Copy activity.

• Eventstreams: Designed for real-time ingestion, transformation and processing of streaming data. Supports low-latency pipelines, schema management, and routing to destinations like Eventhouse, Lakehouse, Activator and Custom Endpoints supporting (AMQP, Kafka and HTTP endpoints).

Detailed feature comparison

The following tables compare the full capabilities of each data movement option. Use this section when you need to evaluate specific features for your scenario.

Scenarios

Review these scenarios to help you choose which data movement strategy works best for your needs.

Scenario 1

James is a finance manager at an insurance company. His team uses Azure SQL Database to track policy data, claims, and customer information across multiple business units. The executive team wants to create real-time dashboards for business performance monitoring, but James can't allow analytics queries to slow down the operational systems that process thousands of daily transactions.

James already has ETL processing in place, and his team needs the processed, gold-tier data available in Fabric for executive reporting. He doesn't want to manage scheduling, configure incremental loads, or worry about table selection - he needs everything mirrored automatically. Since this is for reporting only, having the data in a read-only format in OneLake works perfectly. The solution also needs to be cost-effective since it's coming out of his department budget.

James chooses Mirroring. Mirroring provides the CDC-based continuous replication he needs, automatically handling all tables without any configuration. The simple setup means he doesn't need technical expertise, and the free cost fits his budget. The read-only tabular format in OneLake gives his team the analytics access they need without affecting operational performance.

Scenario 2

Lisa is a business analyst at a logistics company. She needs to ingest raw shipment data from multiple Snowflake databases into Fabric Lakehouse tables for supply chain analysis. The data includes both historical records for the initial load and new shipments that arrive throughout the day. Lisa wants to run this process on a custom schedule - every 4 hours during business hours.

Since Lisa is bringing in bronze-tier raw data, she knows she'll quickly need transformations, schema mapping, and scheduling control. She needs to select specific tables from each Snowflake instance, map columns to standardized names, and use upsert behavior to handle updates to existing shipment records. She also wants advanced monitoring to track data quality and processing performance.

Lisa selects Copy job. Copy job provides the custom scheduling she needs, supports all data sources including Snowflake, and offers the table and column management capabilities for her multi-region setup. The native support for incremental copy with watermark-based detection and upsert behavior lets her handle these requirements without building pipelines.

Scenario 3

David is a senior data engineer at a telecommunications company. He's building a complex data ingestion workflow that needs to extract customer usage data from Oracle using custom SQL queries, apply business transformations, and load it into multiple destinations including both Fabric Warehouse and external systems. The workflow also needs to coordinate with other pipeline activities like data validation and notification steps.

David needs full control over the copy process, including the ability to use user-defined queries to join tables and filter data at the source. He needs advanced and fully customizable configuration options, predictable performance for large data volumes, and the ability to integrate the copy process into broader pipeline orchestration workflows with dependencies and error handling.

David reviews the available options and chooses Copy Activities in Pipelines. This approach gives him the advanced and fully customizable configuration he needs, supports user-defined queries for complex data extraction, and provides the pipeline-based orchestration required for his workflow. The advanced monitoring and auditing capabilities help him track the complex process, while the pipeline framework lets him coordinate copy activities with other data processing steps.

Scenario 4

Ash is a product manager at a telecom company. Her team needs to monitor customer support metrics like call volumes, wait times, and agent performance, in real time to ensure SLA compliance and improve customer satisfaction. The data comes from multiple operational systems including CRM platforms, call center logs, and agent assignment databases, and arrives at high frequency throughout the day.

Ash uses Fabric Eventstreams to ingest and transform this data in motion. She configures streaming connectors to pull data from various sources, applies transformations using the no-code experience, and routes the processed events to Eventhouse for real-time analytics. She integrates Data Activator to trigger alerts and automated workflows when SLA thresholds are breached so she can send notifications to supervisors or adjusting staffing levels dynamically.

The result is a real-time dashboard that updates within seconds, giving Ash's team visibility into live performance metrics and enabling fast, data-driven decisions. This streaming architecture eliminates the latency of batch pipelines and empowers the business to respond instantly to customer needs.

Get started

Now that you have an idea of which data movement strategy to use, you can get started with these resources:

• Get started with Mirroring
• Get started with Eventstreams
• Create a Copy Job
• Create a Copy Activity