Introduction to IoT Architecture

IoT systems stack in 4 layers: sensing gathers, network carries, data processing thinks, application serves: one tank-level reading traced bottom to top makes the whole diagram permanent.

10 min read · 9 cards · 2 checks

Read in: English · हिन्दी · ગુજરાતી


Theory

From a float in a tank to a phone in a pocket

SmartHostel's first real installation: a level sensor in the rooftop tank. The promise: when water drops below 20%, the warden's phone pings and the pump starts by itself.

Between the float bobbing in the tank and the ping in his pocket lies a JOURNEY: measured, transmitted, stored, judged, displayed. Every IoT system organises that journey the same way, and the organisation has a name every exam asks for: the IoT architecture.

Theory

The hostel mess, as architecture

Think of how the mess feeds 200 students: the FARM grows raw vegetables; a TEMPO carries them to the kitchen; the KITCHEN cooks them into meals; the COUNTER serves each student their plate.

IoT stacks identically: sensors grow raw data, the network transports it, processing cooks it into decisions, and the application serves it to humans. Nobody eats raw brinjal from a tempo; nobody reads raw sensor voltages off a wire.

At a glance

The 4-layer IoT architecture

LayerJobSmartHostel example
4. ApplicationUser-facing services: dashboards, alerts, domain appsThe warden's app showing 18% and pinging
3. Data processingStore, filter, analyse, decide (middleware/cloud)A week's readings; the rule: below 20% means alert + pump
2. NetworkCarry data between devices and processingThe hostel WiFi hauling readings to the server
1. Sensing / perceptionSensors measure, actuators act: raw data is bornThe float sensor in the tank; the pump's starter

Theory

Trace one reading up, one command down

Upward (data): the float measures 18% (sensing) → the reading rides WiFi to the server (network) → the rule engine compares 18 against the 20 threshold and decides: alert + refill (processing) → the app shows the red tank icon and pings (application).

Downward (commands): the warden taps Start pump, or the rule fires automatically → the command descends the same layers in reverse → the actuator at the sensing layer switches the pump on.

Data climbs, commands descend: the 2-way traffic is the architecture in motion.

Quiz

In the 4-layer model, WHERE is the decision "18% is below 20%, start the pump" actually made?

  1. The sensing layer: the float sensor decides
  2. The network layer: WiFi routers compare values
  3. The data processing layer: stored readings meet the rule logic there
  4. The application layer: the phone app does all thinking
Show the answer

The data processing layer: stored readings meet the rule logic there

Thinking is the processing layer's whole job: it holds the history, the thresholds and the rules, turning a raw 18 into the decision refill. The sensor (option A) only measures: it does not know 20 is special. The network (option B) is a courier: it carries, never reads. The app (option D) is the serving counter: it DISPLAYS the decision and forwards the warden's taps, and putting all logic in the phone would mean no phone, no pump. One-line exam answer: sensing measures, network carries, processing decides, application presents.

Think first

The 3-layer version in the other textbook

A classmate's guidebook shows only 3 layers: perception, network, application. Is one of you wrong? Work out how the 2 models map onto each other before tapping.

Show the answer

Both are standard: the 3-layer model simply folds data processing into the application layer (or treats it as middleware between network and application). Perception = our sensing layer; network = network; application = processing + presentation together. Exams accept either IF you are consistent and can name the mapping: the safest move is to present the 4-layer model and add one line: some texts merge processing into the application layer, giving a 3-layer view. Knowing WHY models differ (where to draw the thinking boundary) is worth more than either diagram.

Watch out

Architecture-answer slips

Layers as a parts list: the diagram is about the JOURNEY: always trace one reading upward through your answer; examiners reward the trace.

Forgetting the downward direction: commands to actuators descend the same stack; an architecture that only reports is half an architecture.

Putting protocols in the wrong lesson: WiFi, MQTT and friends LIVE at the network layer but their family tour is the NEXT topic: here, name the layer's job, not its zoo.

Theory

Layered thinking is a transferable habit

You have met layer-stacks before and will again: BCA103's hardware-OS-application stack, and BCA502 (Networks) will hand you the OSI model, which is this idea grown to 7 floors. The engineering payoff is always the same: each layer talks only to its neighbours, so any layer can be swapped (WiFi for 4G, one cloud for another) without rebuilding the rest. Next lesson we open layer 2's toolbox: the physical design of IoT and its protocol families.

Summary

Key takeaways

  • IoT architecture: sensing/perception, network, data processing, application: bottom to top.
  • Sensing measures and acts; network carries; processing stores and decides; application presents.
  • Data climbs the stack, commands descend it: trace both directions in answers.
  • The 3-layer variant merges processing into application: name the mapping if asked.
  • Decisions live in the processing layer, never in the sensor or the courier.
  • Layer independence lets any floor be swapped without rebuilding the tower.
  • Memory hook: farm, tempo, kitchen, counter.

Study this properly

This page is the lesson to read. In Gri-Learn the same topic is a graded deck: the self-checks are scored and your weak topics are tracked. Free to start.

Start this topic

Already have an account? Sign in

More from Introduction to Internet of Things

Gri-Learn · syllabus-mapped B.C.A. lessons in English, Hindi and Gujarati

Introduction to IoT Architecture · Internet of Things (IoT) (Minor option A) · Gri-Learn