Data flow from device to network to cloud

Key takeaways

You have built the data path from your hardware to the network gateway. But what actually happens inside it once a connection is active? This page traces the complete lifecycle of a single data packet, from the moment a sensor reading is generated until it lands in a cloud application. Understanding this end-to-end flow is what lets you make informed decisions about protocols, power, and performance.

What you'll learn

The four-layer architecture of a cellular IoT system

Every cellular IoT system shares the same four-layer structure:

• Device layer: The physical reality of sensors capturing the environment, the microcontroller processing data, and the modem handling transmission. This is where the analog world meets the digital one.
• Connectivity layer: The radio stack that carries data into the network, including the antenna, the SIM providing identity, and the registration process that moves data into the core network.
• Data processing layer: Where raw information is organized, filtered, stored, and analyzed either at the edge on the device or in the cloud.
• Application layer: The business end with dashboards, API integrations, and alerts, where raw telemetry becomes actionable intelligence.

How a single data packet travels from sensor to cloud

Following one temperature reading reveals five distinct stages of travel:

1. On the device: the sensor takes the reading and the microcontroller formats it into a payload, wrapped by a lightweight application protocol such as MQTT or CoAP running on top of a transport layer (TCP for reliability, UDP for speed).
2. Over the air: the modem transmits the payload over the radio interface (the "Uu" interface) to the nearest cell tower. Depending on the technology (LTE-M, NB-IoT, or 5G) and signal strength, this can be near-instant or take some "air time."
3. Into the core: the tower passes the packet into the Radio Access Network and then the core, where it enters the User Plane and moves through the Serving Gateway (SGW) and PDN Gateway (PGW).
4. User plane vs. control plane: the User Plane carries your actual data, kept completely separate from the Control Plane, which handles signaling such as paging a sleeping device or tracking its location.
5. To the cloud: the packet leaves the PGW for your cloud platform over the public internet or a private connection, where the application parses, stores, and acts on the payload.

Why the user plane and control plane are kept separate

Inside the core network, your temperature data rides the User Plane, the dedicated highway for real traffic The Control Plane independently handles signaling like paging and location tracking. Splitting these paths lets the network manage devices and move data efficiently without the two interfering with each other.

How data transmission drives battery life

Data transmission is the single largest power consumer in a cellular IoT device. Each time the modem turns on it draws orders of magnitude more current than the sleeping microcontroller, so the data-flow architecture directly shapes battery life. Sending one small packet per hour over NB-IoT with Power Saving Mode between transmissions can yield ten-plus years on two AA batteries, while continuous streams or frequent acknowledgments can drain the same battery in months. Protocol choice matters here too: CoAP over UDP requires fewer round trips than MQTT over TCP, meaning shorter radio-on time and lower energy per message across millions of transmissions.

Lesson 4.3 Self-assessment

Frequently asked questions