liveBook · Manning

Network

As explained in Think Distributed Systems

The communication infrastructure through which distributed-system components send and receive messages.

A network is the communication infrastructure through which concurrent components send and receive messages. It has its own local state, including messages in flight, while each component retains exclusive access to its own local state. section 1.0 section 1.4

A useful mental model treats the network as a participant in system behavior. Components and the network take discrete steps: components perform local computation or send and receive messages, while the network may carry, delay, or fail to forward a message. A component sees only its own state and its channel to the network, not the system’s complete state. section 1.4 section 1.5.2

Distributed Systems Incorporated

Imagine an office whose rooms represent components and whose pneumatic tubes represent the network. A room sends a message into the tubes addressed to another room, but the mailroom may lose, duplicate, or reorder it. The analogy makes message delivery and network failures concrete without treating components as directly connected. section 1.6 section 1.5.2

Figure 8.5 The network as point-to-point communication links between components
A diagram of a algorithm

AI-generated content may be incorrect.

Use the point-to-point network view to visualize communication links and network partitions between components. section 8.4.1

Network behavior shapes correctness, scalability, and reliability. In replication, message loss is modeled as a temporary, intermediate, or permanent failure of one or more communication links, producing a network partition. section 1.9 section 8.4.1

How Think Distributed Systems covers it

How deep it goes
Models local state, message flow, discrete steps, indirect delivery, and partitions, but stops short of implementation details.
Examples
Uses pneumatic tubes, in-flight messages, indirect addressing, and point-to-point links to illustrate communication and failure.
What you can do after
Evaluate distributed-system behavior by separating component state from network behavior and recognizing message loss as partition.
How it connects
Connects distributed systems with components, messages, local state, concurrency, replication, and network partitions.

Also explained in

Compare how 3 books explain network →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Think Distributed Systems. The section links point to the parts of the book it draws on.