ActiveMQ - Journey to 1 Billion msg/day
summary
Claims ability to run a 1 billion messages/day fabric by adding ActiveMQ plugins for multi-tenancy via credential namespaces so identical queue names route to different projects and a destination sharing feature that publishes streams without handing out credentials. Uses a heterogeneous topology: DNS discovery brokers, gateway brokers, master broker that ingests advisories into a TinkerPop graph for advisories and Elasticsearch for metrics. Deployment pipeline uses Docker, Ansible and Rancher, brokers can be ephemeral (example 8 hour lifetime) and one client produced ~1.25M messages in a few hours.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:52Innovation Challenge
- 3:43the journey begins...
- 6:51ActiveMQ Competitors
- 8:23Service Oriented Architecture
- 10:01Service connectivity
- 12:02Open Source ESB Distros
- 12:58Active MQ Features
- 14:04Client Bindings
- 14:37Broker Topologies
- 28:42Identity Management
- 33:57Governance
- 34:51Administration
- 35:39Adaptive Topology
- 38:20Application Delivery Pipeline
- 39:57Conclusion
related talks
22mLessons Learned from Scaling RabbitMQ in a High-Growth (...)
48mA Beginner's Guide to Inter-Service Messaging
24mAdvanced RabbitMQ Strategies: A case study
19mBuilding a Distributed Data Ingestion System with RabbitMQ
41mForget about HTTP
32mBuilding distributed system for retail with RabbitMQ
53mKeynote: AMQP Ten Years After
40mScaling RabbitMQ