Communication Patterns in Distributed Systems

Choose how updates reach clients by considering who initiates delivery, how long connections remain open, and how clients recover missed events. This lesson compares polling, WebSockets, webhooks, and server-sent events, then explains what heartbeats and timeouts can tell you.

A chat window and a monthly report have different freshness requirements. Sending every update immediately may be useful for one and unnecessary for the other.

A delayed update can result from queueing, slow processing, network loss, or the communication pattern itself. Measure where the delay occurs before changing protocols.

Start by asking how stale the displayed information may be and whether clients need to send frequent messages back.

Connection count, message volume, reconnection behavior, and slow consumers matter alongside latency. A protocol choice should account for each of them.

With that context in mind, let’s now explore the two fundamental communication patterns, pull and push.

Push and pull interaction models

At the highest level, communication follows two fundamental interaction models: pull, where the client asks for updates, and push, where the server delivers them.