Overview: Developing a new I/O connector

A guide for users who need to connect to a data store that isn’t supported by the Built-in I/O connectors

To connect to a data store that isn’t supported by Beam’s existing I/O connectors, you must create a custom I/O connector. A connector usually consists of a source and a sink. All Beam sources and sinks are composite transforms; however, the implementation of your custom I/O depends on your use case. Here are the recommended steps to get started:

  1. Read this overview and choose your implementation. You can email the Beam dev mailing list with any questions you might have. In addition, you can check if anyone else is working on the same I/O connector.

  2. If you plan to contribute your I/O connector to the Beam community, see the Apache Beam contribution guide.

  3. Read the PTransform style guide for additional style guide recommendations.

Sources

For bounded (batch) sources, there are currently two options for creating a Beam source:

  1. Use ParDo and GroupByKey.

  2. Use the Source interface and extend the BoundedSource abstract subclass.

ParDo is the recommended option, as implementing a Source can be tricky. See When to use the Source interface for a list of some use cases where you might want to use a Source (such as dynamic work rebalancing).

(Java only) For unbounded (streaming) sources, you must use the Source interface and extend the UnboundedSource abstract subclass. UnboundedSource supports features that are useful for streaming pipelines, such as checkpointing.

Splittable DoFn is a new sources framework that is under development and will replace the other options for developing bounded and unbounded sources. For more information, see the roadmap for multi-SDK connector efforts.

When to use the Source interface

If you are not sure whether to use Source, feel free to email the Beam dev mailing list and we can discuss the specific pros and cons of your case.

In some cases, implementing a Source might be necessary or result in better performance:

For example, if you’d like to read from a new file format that contains many records per file, or if you’d like to read from a key-value store that supports read operations in sorted key order.

Using ParDo and GroupByKey

For data stores or file types where the data can be read in parallel, you can think of the process as a mini-pipeline. This often consists of two steps:

  1. Splitting the data into parts to be read in parallel

  2. Reading from each of those parts

Each of those steps will be a ParDo, with a GroupByKey in between. The GroupByKey is an implementation detail, but for most runners GroupByKey allows the runner to use different numbers of workers in some situations:

In addition, GroupByKey also allows dynamic work rebalancing to happen on runners that support the feature.

Here are some examples of read transform implementations that use the “reading as a mini-pipeline” model when data can be read in parallel:

For data stores or files where reading cannot occur in parallel, reading is a simple task that can be accomplished with a single ParDo+GroupByKey. For example:

Sinks

To create a Beam sink, we recommend that you use a ParDo that writes the received records to the data store. To develop more complex sinks (for example, to support data de-duplication when failures are retried by a runner), use ParDo, GroupByKey, and other available Beam transforms.

For file-based sinks, you can use the FileBasedSink abstraction that is provided by both the Java and Python SDKs. See our language specific implementation guides for more details: