Trainer: Robert Virding

Target Audience: This course is aimed at experienced Erlang Software Developers and Designers who need to understand Behaviours.

Prerequisites: Existing experience using Sequential and Concurrent Programming with Erlang on projects.

Objectives:

• Use existing Design Patterns supported by Erlang and OTP.

• Apply Generic Behaviours, Finite State Machines and Event handler Patterns.

• Use the Supervisor and Applicaton Behaviours Patterns.

• Write your own Design Patterns.

• Structure large Erlang based systems.

Goal: Design fault-tolerant systems.

Duration: Three days

Registration: 08:30 on 11 June 

Venue: 

Introduction

This section introduces the need and advantages of using middleware in the developmentof massively concurrent, distributed, fault tolerant real time systems in Erlang. It looks at Erlang's evolution, and how it helped form OTP. Further this gives an overview of the components that make up OTP. They are Erlang, a set of behaviours, and rules and design principles to use them. OTP comes with many ready built applications and tools useful for large scale software design. They are introduced in this section.

Behaviours

Erlang processes display similar properties. In the Open Telecom platforms, these properties have been abstracted in a group of generic library modules, also called OTP behaviours. This section introduces the concept of behaviours, and through examples of abstraction, shows their need and advantages.

Generic Servers

Generic servers implement the Client-Server behaviours in OTP. This section introduces the most commonly used behaviour in Erlang based applications.

Finite State Machines

The finite state machines behaviour in Erlang is used to implement state transitions in processes based on incoming events. This behaviour is commonly used when implementing protocol stacks.

Supervisors

Supervisors are a behaviour whose only task is to start other Erlang behaviours and monitor them for abnormal termination. This is a vital section in understanding start and restart strategies in Erlang.

Event Handlers

Erlang has ready built event managers. Event handlers are behaviours who subscribe to events sent to specific managers, allowing several actions to take place based on one single event. Event handlers can be changed on the fly, as well as added or deleted from a specific manager.

Applications

This section introduces the application behaviour, used for packaging Erlang resources. Resources can vary from libraries to process clusters, and can be configured to run on a single processor or be distributed on a set of nodes.

Special Processes

There will be times where we want to implement our own behaviours, or for efficiency reason, use simple Erlang processes. This section looks behind the scenes on how behaviours are implemented, and explains how to implement your own.

System Principles

The System Principles section describes how Erlang applications are coupled together in a release and started as a whole. Sections include creating release files, bundling up the software including the virtual machine, and running Erlang on target and embedded hosts.