Skip to main content

Nodes, Topics, and Services - Summary

Quick Reference: Key concepts from this lesson

Core Concept​

ROS2 nodes are independent processes that communicate through topics (publish/subscribe for continuous data) and services (request/response for commands), enabling modular robot architectures where components coordinate without tight coupling.

Key Points​

  • Nodes as Independent Processes: Each node runs separately with isolated memory and single responsibility, providing fault isolation—if one node crashes, others continue operating.

  • Topics for Continuous Data (Pub/Sub): Publishers broadcast messages asynchronously to named topics; multiple subscribers receive data without knowing the source. Perfect for sensor streams (cameras, IMU, encoders).

  • Services for Commands (Request/Response): Synchronous request-response pattern where client waits for confirmation. Use for discrete actions requiring acknowledgment (queries, calibrations, movement commands).

  • Quality of Service (QoS) Policies: Tune reliability vs. performance tradeoffs per-topic—reliable delivery for critical commands, best-effort for high-frequency sensor data tolerating loss.

  • Pattern Selection Impacts System Behavior: Using topics for commands causes lost messages; using services for continuous data blocks unnecessarily. Choose based on: continuous vs. discrete, acknowledgment needs, timing requirements.

When to Use​

Topics (Pub/Sub):

  • Continuous sensor data streams (camera, LiDAR, IMU, encoders)
  • State updates published at regular intervals
  • One-to-many or many-to-many communication
  • Data where latest value matters more than guaranteed delivery

Services (Request/Response):

  • Discrete commands requiring confirmation (move to pose, grasp object)
  • Queries for current state (get position, check battery level)
  • Operations completing quickly (< 1 second)
  • One-to-one transactions needing acknowledgment

Common Patterns​

  • Sensor nodes publish continuously to topics at fixed rates without blocking
  • Processing nodes subscribe to multiple topics, fuse data, publish results
  • Command nodes use services to request actions and wait for completion
  • Status topics use reliable QoS with transient-local durability for late-joiners
  • High-frequency topics use best-effort QoS to tolerate loss for lower latency
  • Previous Lesson: ROS2 Fundamentals - what ROS2 is and why modularity matters
  • Next Lesson: Python rclpy Bridge - implementing these patterns in actual Python code
  • Builds To: URDF robot modeling (Lesson 4), Capstone project with full system integration