CIoTP flashcards
CIoTP · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Latency budget formula
- End-to-end latency equals sensor sampling period plus network transit time across hops plus processing and queueing plus actuation time. Use worst-case values for safety-critical control paths.
- Battery life estimate
- Divide battery capacity in mAh by average current in mA, where average current weights active and sleep currents by their duty cycle. Derate the result for temperature and cell aging.
- Data volume sizing
- Bytes per day equals payload bytes times samples per day times device count. This figure sizes backhaul bandwidth, storage retention, and cloud ingest charges during planning.
- Total cost of ownership
- Sum hardware, installation labor, connectivity subscriptions, cloud and storage fees, and maintenance visits over the depreciation period. Retrofit installation often exceeds device hardware cost.
- Requirement triage
- Separate functional requirements such as sensing rate and actuation latency from non-functional constraints such as power, cost, size, and environmental rating before selecting any hardware.
- IoT risk register
- A list of identified risks such as single gateway failure, firmware end-of-life, spectrum congestion, and physical tampering, each with a mitigation and a named owner.
- MQTT QoS levels
- QoS 0 is at-most-once with no acknowledgment. QoS 1 is at-least-once and may duplicate. QoS 2 is exactly-once with extra handshake overhead. Choose per message criticality.
- MQTT retained message
- A message the broker stores on a topic and delivers immediately to any new subscriber. It provides last known state without waiting for the next device publish.
- MQTT last will
- A message the broker publishes on a device's behalf when the connection drops ungracefully. Commonly used to flag offline status to monitoring subscribers.
- CoAP protocol facts
- CoAP runs over UDP, uses confirmable and non-confirmable messages, supports an observe option for push notifications and block-wise transfer for large payloads, and secures with DTLS.
- Protocol selection rule
- Use MQTT for many-to-broker telemetry over TCP, CoAP for constrained UDP devices, LoRaWAN or NB-IoT for long-range low-power links, and BLE or Thread for short-range use.
- Edge versus cloud split
- Place safety interlocks, closed-loop control, and bandwidth-heavy filtering at the edge. Place long-term analytics, cross-site correlation, and model training in the cloud.
- IoT security baseline
- Unique per-device credentials, mutual TLS or DTLS, encryption in transit and at rest, signed OTA images, least-privilege broker ACLs, and segmented IoT network VLANs.
- Topic hierarchy design
- Structure topics as site, building, device type, device ID, then measurement. Design subscriptions to avoid broad wildcards that force the broker to fan out excessively.
- Idempotent message handling
- Handlers tolerate duplicate and out-of-order delivery by carrying device timestamp and sequence number in the payload, so reprocessing the same message yields the same state.
- Firmware partition layout
- A typical layout has a bootloader with secure boot verification, an application slot, an OTA staging slot, and a persistent configuration store using wear leveling.
- Nyquist sampling rule
- Sample faster than twice the highest frequency of interest in the signal. Undersampling aliases higher frequencies into the measured band and corrupts derived analytics.
- Compact payload encoding
- CBOR, protobuf, or packed binary structs reduce payload size versus JSON on constrained links. A JSON float can cost ten or more bytes against four bytes binary.
- Local buffering policy
- When connectivity drops, queue telemetry to non-volatile storage in a circular buffer with a defined drop policy, usually oldest-first, to bound storage use.
- Command idempotency token
- Downlink commands carry a unique command ID and expiry so that retries do not double-actuate a relay or valve after a lost acknowledgment.
- Hardware-in-the-loop testing
- Tests that run firmware on real or simulated hardware against sensor drivers and protocol peers, exercising timing, bus behavior, and recovery that host unit tests cannot cover.
- Time synchronization approach
- Devices sync via NTP or GPS, store timestamps in UTC with a monotonic sequence number, and let the platform convert to local time for display.
- Staged firmware rollout
- Deploy signed firmware to canary devices first, verify health checks, then expand in cohorts while retaining the previous known-good image for rollback.
- Device provisioning sequence
- Create identity, inject credentials into a secure element or encrypted store, join the network, register the device, then assign it to a site or tenant.
- Commissioning acceptance checks
- Verify RF signal strength at the mounted position, supply voltage, time sync, round-trip latency to the broker, and successful publish and subscribe on each required topic.
- Gateway placement rule
- Mount gateways above obstructions with antenna clearance and confirm coverage using a post-installation site survey rather than trusting the original design model.
- Zero-touch enrollment
- Devices self-register using a bootstrap claim certificate or manufacturer serial allowlist, and the bootstrap credential is revoked after first successful enrollment.
- IoT network segmentation
- Place IoT devices and management interfaces on separate VLANs, allowing only required outbound ports such as 8883 for MQTT over TLS and 5684 for CoAP over DTLS.
- Pilot before scale
- Commission a small representative site, compare measured latency and power against the plan, then expand in batches each gated by an acceptance checklist.
- Dead-man heartbeat alert
- Alert when expected data stops arriving within a timeout window, rather than waiting for an error code, because silent device failure is the most common IoT fault.
- Credential rotation and revocation
- Rotate certificates and keys on a defined schedule, revoke compromised identities in both registry and broker ACLs, and audit for default or shared credentials.
- Telemetry retention tiers
- Keep raw telemetry for a short window and aggregated rollups for the long term. Tiering controls storage cost while preserving trend analysis.