Stateful Functions 2.0 is out.

https://flink.apache.org/news/2020/04/07/release-statefun-2.0.0.html

It allows you to use @ApacheFlink as the base of a *stateful serverless* platform. Flink takes care of your state and messaging, while your application runs as a stateless #kubernetes deployment, or as FaaS functions.
You can run Functions on a platform like @awscloud Lambda, and get out-of-the-box consistent and scalable state, and fast messaging.

Or imaging having your Flink application code run as a separate K8s deployment, supporting rapid scaling and zero-downtime upgrades.
Stateful Functions implements a form of "role reversal" between the DB and the App.

A traditional DB reacts to the application. In StateFun, the DB handles the messaging and the application "reacts to the database."

We call this "event-driven database" for lack of a better term
Stateful Functions implements a pattern of "logical co-location" of state and compute, while enabling "physical separation".

State and compute are handled together in the same way as in stream processing, but the actual execution of the functions can be in a different tier.
This opens up an interesting new architecture for event-driven applications. Something between #streamprocessing, actor programming, and serverless application platforms.

More details for the curious - check the website/docs:

https://statefun.io/ 

https://ci.apache.org/projects/flink/flink-statefun-docs-release-2.0/
You can follow @StephanEwen.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: