It's exciting in the sense that you can apply all these open-source projects to create a very slick lightweight relational database that can replicate changes globally across unlimited instances by using a Kafka topic as the WAL log.

Don't actually use this in production. Use a real database that does what you want, and use the native replication features. If you need to replicate to heterogenous or alternative sources or just want a stream of all changes then you can use Kafka with any number of existing database connectors that automatically read and export changes.

Ever wanted a database with all the operational complexity of Kafka, and more? Well now you can have it!

It’s exciting as someone more in the SRE side, because you’d need to pay be an obscene amount to support this.

Isn't that "side" just a bunch of clicks on a wizard on some cloud provider?

It's like the bulked-up Cartman of key-value stores.

Edit: Nope, it's a relational DB

There are companies that fit the following operational niche:

* Need to run Kafka (for domain-specific reasons)

* Need a relational database, but not necessarily a large or performant one

A project that lets you "buy a Kafka cluster, get the relational database free" is strictly simpler and cheaper to operate (in theory...) compared to operating both a Kafka cluster and a traditional relational database.

If you want an asymmetric persistence layer where all writes are documents placed in a Kafka topic, and all reads are via relational queries.

It's a distributed, eventually consistent, highly available ledger that can be queried with SQL.

Is not eventually consistent, it uses https://omid.incubator.apache.org/.

It uses kafka as an operational primitive. Kafka is somewhat complicated but it also solves real problems. I can see a fast, distributed db built on top of kafka.

There fast distributed DBs and they are generally built on top of distributed ordered key-value stores. Many of them do not require to run external components like zookeeper. What is the advantage of running one on top of Kafka?

The main benefit I have heard is that some people already have Kafka.

in what sense is it a benefit?