2 IoT Domains: Shared Patterns
More
Depth: Practitioner
Progress saves as you read. Mark sections complete as you go; your progress appears in the sidebar.
- Podcast audio coming soon
- Character narration audio coming soon
2.1 The Decision at Stake
An application domain is an industry or sector where connected sensors and devices create value. A farm, ward, factory, and city can share sensing, tracking, control, and alert patterns. What makes those patterns useful is the decision they improve and the person or service responsible for acting.
Picture a heart monitor and a city bin sensor. Both send readings, but a late heart warning and a late bin report do not cause the same harm. Their users, laws, power needs, and support paths differ too. A false heart alert may cause fear and care work; a missed alert may cause direct harm. A false bin alert may waste a trip, while a late bin report may be acceptable until waste blocks a street.
The design problem is deciding which assumptions can travel with a shared sensor pattern. Start with the outcome, the action, its time limit, and the cost of a wrong result. Then ask what happens when a link fails, a reading is wrong, an update fails, or the deployment grows. A domain label alone cannot answer those questions.
This first part compares the Five Pillars framework and its 14 sectors, relates sensor and communication needs to each setting, and tests when patterns transfer. It also gives business leaders a way to compare value without hiding operational owners. The approaches begin with a broad domain map, then narrow to the evidence needed for a particular service.
KEY CONCEPTS
- Domain claim
State the job before borrowing a sensor pattern.
- Transfer test
Check whether timing, power, and trust assumptions survive the move.
- Service owner
Name who responds when the device reports a fault.
2.2 Part Objectives
- Identify shared sensing, tracking, control, and alert patterns.
- Compare domain value without hiding operational owners.
2.3 Use the Pillars to Ask Better Questions
The Five Pillars framework groups IoT applications by their purpose. Use it to locate a problem before choosing sensors or connectivity. The chapter’s headline examples illustrate different kinds of value and requirements; they do not establish what every deployment in a pillar will achieve.
| Pillar | Domain emphasis | Headline example in this domain map |
|---|---|---|
| SUSTAIN | Cities and shared infrastructure | 30% energy savings in smart buildings. |
| MOVE | Transport, location, and routing | Sub-10 ms latency for autonomous vehicles. |
| HEAL | Healthcare and clinical workflows | 99.9% reliability. |
| FEED | Agriculture and field operations | 40% water savings in precision agriculture. |
| MAKE | Manufacturing and production uptime | 25% less downtime. |
The numbers point to different design questions. Latency spans under 10 ms for autonomous vehicles to hours for agricultural sensors. Power can range from a mains-powered industrial installation, without a battery-life limit, to a LoRaWAN parking sensor designed for 10+ years while transmitting 10 bytes/hour. Match a sensor and communication path to the required response and power budget before treating either example as a template.
A hospital-department analogy helps distinguish those timescales: transportation resembles the Emergency Room, where milliseconds matter; a smart home resembles a General Ward making hourly checks, where minutes may be acceptable; agriculture resembles a Lab running daily tests, where hours may be acceptable. This is an analogy for different requirements, rather than a timing specification for every system in those sectors. Similar tools can still demand different speed, reliability, and precision.
The settings also change what can be installed and maintained. A home has residents and a local network. A farm has land, weather, animals, distance, seasonality, weak power, and field maintenance. A plant has machines, shifts, safe states, old equipment, plant networks, and maintenance ownership. A road adds motion, weather, and many owners. HEAL adds clinical workflow, safety, and privacy obligations.
These differences explain why a smart-parking sensor cannot simply become a patient heart monitor. They also help you choose a project, identify the skills a career needs, or ask useful questions when evaluating a product. For a more detailed comparison, use the domain-requirements figure in Part 2. The companion route, IoT Application Domains: Requirements and Navigation, provides another entry into the detailed chapters.
2.4 Test What Transfers Between Domains
The pillars help locate a use case; a transfer test checks whether a pattern’s assumptions still hold in its new setting. Two health systems may need different architectures, while a farm and a factory may share a useful pattern when their hard limits match.
Predictive maintenance shows the opportunity. The pattern of sensor data, anomaly detection, and preventive action used in manufacturing (MAKE) can also serve medical-equipment monitoring (HEAL). The underlying sequence transfers, but timing, power, trust, and responsibility still need review. Copying the sequence does not establish that its original thresholds or operating arrangements are suitable.
The comparison extends beyond water sensing. A building occupancy estimate can be probabilistic if it saves energy without making rooms unusable. A cold-chain vaccine monitor needs audit history, calibration evidence, and escalation before product quality is questioned. A factory vibration model can help predict a maintenance window, yet damage trust if its alarms stop production without root-cause evidence. The same sensor family may therefore need different sampling rates, retention, interface language, and support arrangements.
A smart-parking sensor, insulin refrigerator monitor, irrigation valve, and factory vibration sensor may all send telemetry, meaning reported measurements or state. Their operating context still decides whether those reports are timely, credible, and useful. Follow the response all the way to the person or service that can act.
2.5 Connect the Decision to Business Value
A transferable pattern needs an operating case as well as a technical fit. Across the 14 domain examples, IoT creates different kinds of measurable value, including the smart-building energy and precision-agriculture water savings above. Operational efficiency, new revenue from data-driven services, and improved customer experiences can give early adopters a competitive advantage. The investment decision depends on whether those benefits occur in the proposed service.
Start where operations generate data that could improve decisions, as in manufacturing, logistics, or facilities. Other reasons to investigate IoT include costly or error-prone manual inspections, connected products that improve customer experience, and regulatory requirements for continuous monitoring and reporting. Gathering readings is useful only when it changes an action or supplies required evidence.
The chapter’s investment framework includes four dimensions:
- Initial investment: Sensors, connectivity, platform, and integration typically range from $25,000 to more than $500,000.
- Operating cost: Platform fees, connectivity, and maintenance commonly range from $1,000 to $25,000 per month.
- Payback: The timeline depends on domain complexity and often falls between 6 and 36 months.
- Risk: Consumer and facilities deployments are usually lower risk, while industrial, healthcare, and transportation deployments carry higher safety, reliability, and compliance exposure.
Use these ranges to frame questions, then measure the proposed service’s value and lifecycle costs. Ask who pays for the device, data, site visits, updates, and final removal. A useful pilot may still fail as a lasting or fair service. A payback period borrowed from another domain cannot establish the case for replacing this team’s manual inspections.
2.6 Build a Reviewable Domain Record
The business case and transfer test meet in a record of the proposed service. Start with one outcome card per domain. Name the person, animal, place, asset, or service that benefits, the decision that changes, and the time in which action still helps. If no action can be named, the data job is not yet clear. Return to the heart monitor and city bin: record their different harms and response owners before choosing a platform.
Write the measurement claim next. State the physical event, unit, range, rate, time, and known error. Add the point where a value is checked, and keep raw and changed forms linked. A shared sensor part does not give a measurement the same meaning in every domain. This record makes it possible to question the evidence behind an alert rather than accepting a plausible dashboard.
Trace the action path through the device, local rule, gateway, network, store, service, human, and final act. Put an owner beside every step. Mark what must stay local, what may wait, and what must never happen from old or doubtful data. Identify who receives an alert, who can act, and what happens when the system is wrong. Sensing, communication, storage, and support should follow from that path.
Group requirements under time, harm, data, place, energy, scale, privacy, security, and support. Give each group a hard need or a measured range. Before choosing sensors, identify enclosure needs, radio path, installation access, available power, and maintenance interval. These fields let two domains be compared without forcing them into a single architecture.
Test the normal path, then remove a sensor, lose the link, delay the data, restart the device, fill a queue, and change the user. Include a failed update and an increase to ten times more units. For every case, keep the safe result and repair owner in the record. Missing, stale, wrong, or offline devices must have an explicit outcome, because each can interrupt the intended decision differently.
Review people as carefully as hardware. Ask who is watched, who is missing, who may say no, who can fix a record, and what data must stay private. The same connected capability earns trust in a city, clinic, farm, factory, store, or home only when it fits that setting’s users, constraints, regulations, and failure consequences.
Choose the smallest first test that could break the claim. Use real people, places, and fault states where safe, and set the pass rule before testing. Keep misses and hard cases. A clean demonstration cannot prove fitness for a full town, ward, plant, or fleet. Selecting technology before understanding latency, reliability, power budget, data volume, and regulatory compliance puts that fitness at risk.
End the record with a narrow release statement. Name the domain, users, sites, devices, action, time limits, failure cases, and owners it covers. Add a date and the change that forces another review. The domain name begins this record; the evidence establishes its scope. Read the following chapters as comparison cases, watching how each shifts power, connectivity, data quality, standards, human workflow, and liability.
Summary
The Five Pillars organize 14 IoT sectors, but each service needs its own timing, harm, power, privacy, and support requirements. The heart monitor and city bin show why similar readings can demand different responses. The home leak alarm and park soil sensor show how a shared sensing pattern can require different behavior during the same outage.
Reuse begins with a transfer test. A reviewable domain record connects the measurement to a decision, names the response owner, tests failure cases, and limits the release claim to the evidence collected. The business case adds operating value, lifecycle cost, payback, and risk.
Start with the decision and its consequences, then reuse only the parts of a pattern whose assumptions survive the move.
Carry that record into IoT Domains: Constraints and Collection Paths, which begins with Compare Domains by Constraints.

