Let Me Just Say It
In 2022, I cannot find a compelling technical reason to choose Azure Data Factory over Synapse Pipelines for a new project, or vice versa. The feature parity is real. The Microsoft differentiation story — different tools for different use cases — is becoming marketing rather than architecture.
That's the direct statement. Here's the evidence.
The Feature Parity Breakdown
Let me go through the things that matter in an enterprise orchestration platform and tell you where each one stands.
Activity Types
ADF: Copy Activity, Data Flow, Notebook (Databricks or Synapse), Stored Procedure, Azure Function, Web Activity, Until, ForEach, If Condition, Wait, Set Variable, Append Variable, Lookup, Get Metadata, Filter, Switch, Fail, Delete, Validation, Execute Pipeline, and more.
Synapse Pipelines: identical list. They share the same activity model. Same JSON schema.
Connector Library
ADF: 100+ connectors across Azure services, databases, SaaS platforms, file formats, and cloud storage.
Synapse Pipelines: same connector library. If a connector exists in ADF, it exists in Synapse Pipelines. The underlying connector infrastructure is shared.
Expression Language
ADF: the expression language for dynamic content — @pipeline().parameters.tableName, @concat(), @formatDateTime(), @json(), the full set.
Synapse Pipelines: identical. Same functions, same syntax, same behavior. If you write an expression in ADF, it runs unmodified in Synapse Pipelines.
Parameterization Model
ADF: pipeline parameters, dataset parameters, linked service parameters, global parameters, trigger parameters.
Synapse Pipelines: same model, same layers. The metadata-driven framework pattern I've been building for years works identically in both environments.
Integration Runtimes
ADF: Azure IR (managed, auto-scaling), Self-Hosted IR (on-premises connectivity), Azure-SSIS IR (SSIS package execution).
Synapse Pipelines: Azure IR and Self-Hosted IR. No Azure-SSIS IR.
This is the one meaningful difference. If you need to run SSIS packages in managed cloud infrastructure, ADF is your product. Synapse Pipelines doesn't have SSIS-IR and shows no signs of getting it.
Git Integration
ADF: native Azure DevOps or GitHub integration. Branch-based development, PR workflow, adf_publish branch for ARM template generation.
Synapse Pipelines: same model. Synapse workspace connects to Azure DevOps or GitHub, the same branch-based development pattern applies, the publish mechanism is analogous. The git configuration UI is slightly different, but the workflow is the same.
CI/CD
ADF: ARM template deployment via Azure DevOps, pre/post trigger scripts, environment parameter files, automated publish via npm package.
Synapse Pipelines: same ARM template approach. The Synapse workspace artifacts deploy via ARM the same way. The CI/CD pipeline structure is nearly identical.
Managed Virtual Network
ADF: Managed VNet with private endpoints for source and sink connections.
Synapse Pipelines: Synapse Managed VNet with the same private endpoint model. Same security capability, different UI.
What Synapse Pipelines Adds
There are scenarios where Synapse Pipelines has a genuine advantage over standalone ADF:
Co-location with Synapse SQL and Spark pools. If you're using Synapse Dedicated SQL Pools or Synapse Spark Pools as your compute, having the pipelines in the same workspace simplifies linked service configuration and reduces network hops. Not a massive difference, but real.
Native Synapse Link integration. Synapse Link — the near-real-time analytical replica capability from Cosmos DB, Dataverse, and SQL Server — is natively integrated with Synapse Pipelines. ADF can work with Synapse Link outputs, but it's not as seamless.
The unified workspace experience. Synapse Studio puts pipelines, notebooks, SQL scripts, and Power BI reports in one interface. For teams that work across all of these, that integration has genuine value.
What ADF Has That Synapse Pipelines Doesn't
One thing: the Azure-SSIS Integration Runtime.
If you're lifting and shifting legacy SSIS packages to managed cloud infrastructure, ADF is your only path in the Microsoft stack. Synapse Pipelines cannot run SSIS packages. End of list.
The Strategic Ambiguity
The harder question is the long-term one. Are these two products being maintained as distinct long-term, or is consolidation coming? Microsoft's messaging — "ADF for enterprise data integration teams, Synapse for analytics co-location" — describes a differentiation that the feature matrix doesn't support.
My bet: consolidation is coming. The unified analytics platform that Microsoft is clearly building will absorb both products under a new name. I've heard Fabric mentioned internally. Whatever it's called, the technology underneath — the JSON pipeline model, the connectors, the expression language — isn't changing.
So my advice: don't make this choice based on which product you think will survive. Make it based on your existing infrastructure and team fit. Both investments are durable. The patterns transfer.
Trust me on this one. I've been through platform transitions before. The skills don't disappear when the product name changes.