🧪 Hacking with Durable Streams + Rivet Actors Actors → the brains & memory Durable Streams → the pipes Electric team → the 🐐 This is VERY experimental – demo code & slides below

@ElectricSQL btw on distributing load — we explicitly designed this to scale broadcast to essentially unlimited numbers — we've tested to 1 million connected clients to on stream — Awesome overview!

@ElectricSQL thanks! seems i should've better articulated broadcast (DS) vs consumer groups/partitions (Kafka)

Source code: Slides:

@ElectricSQL 🫶

@ElectricSQL how would you think about voice actors using streams/actors? Since presumably this would all be over a TCP connection, you'd sacrifice latency, right? Is there a different way you're looking at it?

@ElectricSQL looks really cool

@ElectricSQL love the concept mate! familiar w/ DOs, but not DS diving into your codebase, glad you made it oss let me know if launching you on would help btw

@ElectricSQL Awesome explanation. I'll check the code. Do you plan to release LiveStore examples with durable streams? I know LiveStore has an S2 integration and would love to know how it this new ElectricSQL durable stream fits together with Rivet actors and livestore

@ElectricSQL this looks niiiiiccceee

@ElectricSQL Really cool!

@ElectricSQL This architecture is quite cool. the split between actors as the brains and Durable Streams as the pipes makes ton of sense too. Curious how you’re handling debugging and observability across those streams and actors in practice!

@ElectricSQL i feel like cool things could be built off ye olde iroh is p2p shi was actually consumable. the nouns and verbs composing its lingua franca were terribly chosen.

@ElectricSQL This demo is an instant sub, very cool stuff