GKE and Unity: A matchmaker made in heaven
summary
Unity Matchmaker runs on GKE with Pub/Sub, Firestore, Memorystore and Dataflow, using Terraform and Helm to deploy. It segments players into queues and ordered pools, maps tickets to ticket buckets and schedules match processes on match process hosts via a World Map. The system scales to hundreds of nodes and thousands of CPUs for full-scale tests, allocating hundreds of GBs of RAM, and in lab runs matches a simulated 153 million concurrent users with 95th-percentile time-to-match about 3.3 seconds.
AI-written summary · about
chapters (8)
- 0:00Introduction to GKE and Unity
- 1:03Understanding matchmaking
- 3:20Matchmaking mechanics
- 7:46Unity Matchmaker features
- 14:11Architecture and scaling
- 20:42Leveraging GKE infrastructure
- 24:43Future roadmap
- 26:00Q&A session
related talks
19mHow to make multiplayer matchmaking easier and scalable with Open Match
37mLarge scale multiplayer gaming with open source
42mProcessing Millions of Events per Second with Apache Kafka on Kubernetes
35mKubernetes journey
06mKubernetes Explained the Easy Way, Using Computer Games...
32mScaling Multiplayer Game Servers with Kubernetes
31mAn Adventure in Multi Cloud with Kubernetes and Distributed Data
55mCloud services for gaming