OTP, the Middleware for Concurrent Distributed Scalable Architectures
summary
Claims that OTP cuts code and bugs: reused middleware plus Erlang VM give 4–20× less code than C++ (including OTP) and far fewer defects; teams report automated release pipelines built on OTP principles. Shows concrete tools: gen_server, supervisors, gen_fsm, make_ref, monitors, eunit, common_test, dbg, Observer. Describes supervision trees that isolate failures, explains why release handling is clunky but automatable, and gives practical race-condition and monitoring fixes that demonstrably work in production.
AI-written summary · about
chapters (16)
- 0:00Introduction to scalability
- 1:03Defining system requirements
- 2:04High availability and faults
- 3:07Limitations of languages
- 4:27Introduction to OTP
- 6:19Demystifying OTP name
- 8:38OTP principles and design
- 10:09Libraries and tools
- 11:55Release and debugging tools
- 14:11Advantages of using OTP
- 17:47OTP behaviors and patterns
- 20:27Supervisors and applications
- 22:12Concurrency challenges
- 25:28Error handling and monitoring
- 30:20Fail safe and supervision
- 34:55Conclusion and Q&A
related talks
50mA Peek Inside Erlang's OTP
48m"Handling Millions of Concurrent Users with Erlang OTP"
01h 01mThe Erlang Ecosystem
12mObject Oriented Thinking with Elixir and OTP
39mConcurrency Oriented Programming In A Modern World
04mFrancesco Cesarini interviewed at OSCON 2010
51mThe Erlang VM or How I Started Loving Parallel Progamming
52mKeynote: Pilgrim's Progress to the Promised Land