This document contains the design for a new Actors service for Dapr, which:
- Provides the address for actors running in the cluster, replacing the current Placement service and table distribution with a more efficient and reliable design.
- Manages, stores, and executes actor reminders, in a centralized service.
The new service requires the use of a relational (aka "SQL") database, including Azure SQL (Microsoft SQL Server), PostgreSQL, MySQL, etc… as well as SQLite.
Because the Dapr workflow building block is based on actors, improvements in this space will benefit workflow users too.