Microservice Swarms: Decentralized Discovery and Scaling
summary
Claims a microservice architecture that scales without centralized discovery by using ZeroMQ messaging, JSON payloads over TCP, and lightweight HTTP gateways (mongrel2 style). Services are self-replicating instances that clone when their local queue is overloaded; address tables for all instances flow back in each response so discovery is eventual and decentralized. Heartbeats detect failures, round‑robin push/pull provides load distribution, prototypes (Janus v1) run on cheap nano VMs and handle about 7–8 concurrent instances before design tweaks.
AI-written summary · about
chapters (8)
- 0:00Start
- 15:34As stand-alone web servers
- 18:27Queues
- 27:49Basics
- 28:46Transport
- 30:27Not you grandfather's socket
- 44:23Singletons
- 56:49Winding down
related talks
27mSolving service discovery: Node.js microservices without a registry
26mBuilding Scalable Microservice APIs From The Beginning
27mMicroservices
43mGo micro with Microservices
28mMicroservices – Boyan Dimitrov
03mMicroServices - Let’s Build Some!
48mChallenges in Implementing MicroServices
16mMicroservice Challenges