In IoT, scale changes everything. Ten devices can be configured by hand; ten thousand need automated provisioning, a device registry, bulk updates and monitoring that surfaces problems without a person watching each device. Message volume that a single broker handled comfortably needs clustering. Storage that kept every reading needs retention and downsampling policies. Support costs grow with the fleet unless failure rates and remote diagnostics are addressed.
Scalability is as much about operations and cost as about software architecture. A platform that charges per message may be cheap in a pilot and ruinous at scale, and a radio choice that works in one building may not cover a region. Testing with realistic fleet sizes and failure scenarios before rollout is the practical defense against pilot purgatory.
Part of the IoT Glossary. Spotted an error or a missing term? Tell us.