Search notes:

Apache Kafka

Kafka promises to handle real-time data feeds («Streaming») especially well. It is typically used as

Kafka relies on the TCP/IP protocol.

Terminology

A producer sends messages.

A consumer pulls (reads) messages.

Brokers are the end points/nodes in Kafka clusters.

Messages are organized into topics and stored in partitions.

Messages are stored in key/value pairs in the order or reception (that is: they're time-stamped).

Kafka is written in Scala and Java.

Originally, Kafka was developed by LinkedIn.

With Oracle SQL Access to Kafka (OSAK), Oracle external tables can connect to Kafka brokers and read Kafka records.

Thus, Kafka data can be joined to data in Oracle Database tables.

Note that Kafka data does not have to persist in Oracle Database to be queried.

OSAK also enables you to load Kafka data into tables stored in Oracle Database.