Archive for October, 2021

Processing Apache Avro-serialized Kafka messages with IBM App Connect Enterprise

Monday, October 25th, 2021

IBM App Connect Enterprise (ACE) is a broker for developing and hosting high-throughput, high-scale integrations between a large number of applications and systems, including Apache Kafka.

In this post, I’ll describe how to use App Connect Enterprise to process Kafka messages that were serialized to a stream of bytes using Apache Avro schemas.



Best practice when using Apache Kafka is to define Apache Avro schemas with a definition of the structure of your Kafka messages.

(For more detail about this, see my last post on From bytes to objects: describing Kafka events, or the intro to Avro that I wrote a couple of years ago.)

In this post, I’m assuming that you have embraced Avro, and you have Kafka topics with messages that were serialized using Avro schemas.

Perhaps you used a Java producer with an Avro SerDe that handled the serialization automatically for you.

Or your messages are coming from a Kafka Connect source connector, with an Avro converter that is handling the serialization for you.

Or you are doing the serialization yourself, such as if you’re producing Avro-serialized messages from a Python app.

Now you want to use IBM App Connect Enterprise to develop and host integrations for processing those Kafka messages. But you need App Connect to know how to:

  • retrieve the Avro schemas it needs
  • use the schemas to turn the binary stream of bytes on your Kafka topics into structured objects that are easy for ACE to manipulate and process


From bytes to objects: describing Kafka events

Saturday, October 23rd, 2021

The recording of the talk that Kate Stanley and I gave at Kafka Summit Americas is now available.

Events stored in Kafka are just bytes, this is one of the reasons Kafka is so flexible. But when developing a producer or consumer you want objects, not bytes. Documenting and defining events provides a common way to discuss and agree on an approach to using Kafka. It also informs developers how to consume events without needing access to the developers responsible for producing events.

In our talk, we introduced the most popular formats for documenting events that flow through Kafka, such as AsyncAPI, Avro, CloudEvents, JSON schemas, and Protobuf.

We discussed the differences between the approaches and how to decide on a documentation strategy. Alongside the formats, we also touched on the tooling available for the different approaches. Tools for testing and code generation can make a big difference to your day-to-day developer experience.

The talk was aimed at developers who maybe aren’t already documenting their Kafka events or who wanted to see other approaches.

watch the recording on the Kafka Summit website


Using pitch estimation to play with music in Scratch

Tuesday, October 5th, 2021

I’ve added a pitch extraction machine learning model to Machine Learning for Kids today. In this post, I want to describe the model a little, and suggest a few ways that students could use it.


I started adding pretrained machine learning models to Machine Learning for Kids last year. Although my main focus is still allowing students to create their own machine learning models and make things with them, there are some fun projects that can be made using models that are too complex for students to train by themselves.

imagenet (that I added last Christmas), and the question-answering model (that I added in April) are both good examples of that!

I hope this one will be similarly welcomed!


The new model is a pitch estimation model. Given some audio as input, you can use it to recognize the dominant pitch in sung audio (even if there is background music and noise).