Latency is usually measured as one-way or round-trip time and is shaped by every stage in the path: radio access, sleep cycles, mesh hops, protocol handshakes and cloud processing. A LoRaWAN Class A device may only receive a downlink seconds after it transmits, while a wired industrial bus delivers updates in a fraction of a millisecond.
In IoT the question is less how fast than how predictable. A thermostat tolerates seconds; a robot safety stop or a grid protection relay does not. That is why control logic often moves to edge computing close to the device rather than relying on a cloud round trip, and why 5G low-latency modes and time-sensitive networking matter for factories. Latency and bandwidth are independent: a satellite link can carry a lot of data and still be slow to respond.
Part of the IoT Glossary. Spotted an error or a missing term? Tell us.