The latest version of Diffusion™ contains new features,
performance enhancements and bug fixes.
Note: This is a preview release which adds a number of the new features that will be available in
the full release of Diffusion 6.6.
This preview is supported for production use, except for the Python SDK and the Kafka adapter.
Diffusion now has a SDK for developing clients in the popular Python language.
In this preview release, the Python SDK supports subscribing to topics, and request-response messaging. We plan to expand its capabilities in the full 6.6 release.
For more details, see the Python SDK overview.
Topic view inserts
Topic view inserts are a new addition to the data processing capabilities of topic views. Topic views
enable you to mirror selected source topics to another part of the topic tree, creating reference topics.
With a topic view insert, you can now merge data from topics other than the selected source topic into JSON reference topics.
You can insert whole values, or partial data specified with JSON pointers.
Advanced features include the ability to derive paths from values, and chaining multiple inserts.
For more information, see Insert clause.
This release includes a beta version of a new integrated adapter to interface with Apache Kafka.
The adapter can translate data from Kafka topics to Diffusion topics, and from Diffusion to Kafka.
The new adapter has expanded configuration options, including the ability to configure how imported topics are structured,
and to import topics that match a regular expression.
For more information, see Kafka adapter.
Time series topic enhancements
You can now update a time series topic via the standard topic update API, treating a time series topic
as if it were a single topic with the same event type as the time series. This means that when you update time series topics,
you can now use features like update constraints,
update streams and the addAndSet operation.
In addition, you can now create a time series event with a custom timestamp of your choice, instead of one based on the current
time. You could use this to load historical data into a time series topic, or for testing purposes.
Session filters now support a new IN operator, so you can filter sessions based on a list of properties.
The Android SDK has been updated to take advantage of the latest desugaring support in Java 8+.
Update streams now respect topic conflation settings and can be used to update topics that with DONT_RETAIN_VALUE set to true.
The server will use less CPU when update traffic is very low.