Serving Models Made Easy with Stateful Functions
Stateful Functions is an open source framework that reduces the complexity of building and orchestrating distributed stateful applications at scale. It brings together the benefits of stream processing with Apache Flink and Function-as-a-Service (FaaS), providing a powerful abstraction for the next generation of event-driven architectures.
Stateful Functions applications are a collection of virtual stateful functions that can send arbitrary messages between each other and external systems. The execution can happen on a local JVM, or clusters of many machines. The framework is based on functions with persistent state that can share a pool of resources and interact arbitrarily with strong consistency guarantees. Stateful Functions uses a state-of-the-art runtime built on Apache Flink for distributed coordination, communication and state management.
This framework separates compute and storage differently than the classic two-tier architecture: with Stateful Functions we have an ephemeral state/compute tier, and a simple persistent blob storage tier. This approach eliminates the need to provision additional databases, key-value stores, or message brokers, and effectively offloads application state management from the shoulders of developers.
This workshop will cover the Stateful Functions runtime and the API in enough detail to prepare to you to be immediately productive with it.
During the hands-on sessions of this workshop, we will use Stateful Functions to develop a simple fraud detection service wrapped around a fraud detection model.