3 Agent-based systems

 

This chapter covers

  • How edge computing makes use of agent-based systems
  • Defining agents and using them to perform functions
  • Agent topologies and communication protocols
  • Pitfalls to avoid in using agent-based systems
  • Using edge computing to connect local resources to the cloud

In the previous chapter we introduced an edge-computing testbed that took the form of a single agent that was used to run a small number of functions. Agents are a central concept of any edge-computing framework and your use-case could involve hundreds of them spread across your company network performing thousands of tasks, some of which may work collectively for many different applications; thus it’s important to have a general understanding of how they are set up and function. This chapter presents a bottom-up look at the fundamentals of how agents and agent-based systems are used in edge computing.

3.1   Hiring yourself a secret agent

3.1.1   All systems functional

3.1.2   Getting work done with agents

3.1.3   Let’s appify our edge functions

3.2   Arranging your agents

3.2.1   Simple single-layer topologies

3.2.2   Demo: Building a single-layer topology

3.2.3   Moving to the multi-layer

3.2.4   Demo: Adding another company to your warehouse

3.3   Secret agent communications

3.3.1   Message brokers to bind them all

3.4   When good agents go bad

3.4.1   An agent’s secret identity

3.4.2   Picking up after one another

3.4.3   What’s going on?

3.5   Use case: Bridging agents between home and the cloud

3.6   Summary