Building on developments with version 4.0, PNDA 5.0 adds new features and fixes for deployment that enhance security, add support for Flink and remove constraints for ingestion. These updates will allow for broader applications for users, enabling integration with new components and support the creation of new stream processing applications.
Improved Security with Knox
PNDA 5.0 introduces new security mechanisms with the Knox Application Gateway. Providing improved perimeter security for user services, the latest release now supports configurable authorization. Moreover, through Knox & PAM we now support authentication via LDAP which makes integration in already existing security infrastructure even easier.
A major API change in 5.0 includes modification to the ingestion method, removing the constraint of having only one strict AVRO schema and enabling Gobblin to ingest data in protobuf and other formats.
5.0 brings integration with Apache Flink for an alternative option to Spark. The new integration enables stream processing at event-level granularity as well as very low latency with more advanced windowing in addition to more stream processing capabilities not available with Spark.
PNDA 5.0 will be included in the upcoming Casablanca release of ONAP, following improvements to PNDA related to mirror creation and the added ability for it to be built to a Docker container. Primarily targeted to Openstack deployment, 5.0 also includes updates to Heat templates and the pnda-cli.
As usual, the PNDA 5.0 release includes numerous updates, additions and fixes. Visit the PNDA 5.0 page on Github for the full details.