The BEAM for Developers

This is a full-day, hands-on introduction to the Erlang Runtime System and the BEAM virtual machine. The course focuses on how the runtime works in practice: processes, message passing, memory management, garbage collection, and scheduling. These mechanisms determine how BEAM systems behave under load, during failures, and over long runtimes.

The goal is to give you a correct mental model of the VM so you can reason about performance, reliability, and system behavior without guesswork.

What we cover:

• Thinking in Processes — How to structure systems as many small, isolated processes that communicate through messages. We use the Gnome Village model to reason about responsibilities, failure isolation, and supervision.
• Data Types and Messaging — How data is represented in the BEAM. How process heaps work. What actually happens when messages are sent between processes, including copying, binaries, and references.
• Memory and Garbage Collection — Per-process heaps, generational garbage collection, binary handling, and common sources of memory growth. How to spot and avoid long-term memory problems.
• Scheduling and Concurrency — Reductions, run queues, and SMP schedulers. How the BEAM balances throughput and latency, and why some workloads behave predictably while others do not.
• Debugging and Observability — How to inspect live systems using built-in tools. We work with tracing, profiling, and runtime inspection using tools such as Observer and Recon.
• Patterns and Anti-Patterns — Designs that cause problems in real systems: overloaded mailboxes, large binaries in messages, centralized processes, and hidden bottlenecks.

Format: The day alternates between short theory sections and practical exercises.

Requirements: Laptop with the BEAM and your preferred language installed.