Edge Computing and IoT

Edge computing is a distributed computing paradigm that brings computation and data storage closer to the sources of data. This is in contrast to traditional cloud computing, where all processing occurs in centralized data centers. The primary drivers for edge computing are latency, bandwidth, and reliability. Many applications—such as autonomous vehicles, industrial automation, and real-time analytics—require response times in the order of milliseconds, which is impossible if data must travel to a cloud server hundreds of miles away. Additionally, transmitting massive amounts of sensor data to the cloud consumes bandwidth and incurs costs. Edge computing processes data locally, sending only relevant summaries or alerts to the cloud.

Edge computing relies on edge devices or nodes that sit between the data source (e.g., an IoT sensor) and the cloud. These edge nodes can be dedicated hardware (e.g., gateways, routers with compute capabilities) or virtualized instances running on local servers. The workflow typically follows these steps:

Edge nodes often include a local database (e.g., SQLite, InfluxDB) to store time-series data temporarily. They also maintain a connection to the cloud, which may be intermittent or always-on.

- Edge Devices: These are the physical or virtual devices that perform processing. Examples include Raspberry Pi, NVIDIA Jetson, industrial PCs, and edge-optimized servers. - IoT Gateways: A specific type of edge device that aggregates data from multiple sensors and translates between different protocols (e.g., Zigbee to MQTT). - Protocols: - MQTT: Publish/subscribe protocol over TCP, default port 1883 (unencrypted) or 8883 (TLS). QoS levels: 0 (at most once), 1 (at least once), 2 (exactly once). - CoAP: UDP-based RESTful protocol, default port 5683. Uses confirmable/non-confirmable messages. - HTTP/HTTPS: Common for REST APIs, but heavier than MQTT/CoAP. - AMQP: Advanced Message Queuing Protocol, used in enterprise IoT. - Latency Targets: Edge computing aims for <10 ms round-trip time (RTT) for critical actions, compared to 50-500 ms for cloud. - Bandwidth Savings: Typically 90-99% reduction in data sent to cloud, as only 1-10% of raw data is uploaded.

While edge computing is not a single configuration, administrators often configure edge gateways and IoT devices. Example: Configuring an MQTT broker on an edge device (using Mosquitto):

# Install Mosquitto MQTT broker
sudo apt-get install mosquitto mosquitto-clients

# Edit configuration file /etc/mosquitto/mosquitto.conf
# Example: allow anonymous connections, listen on port 1883
listener 1883
allow_anonymous true

# Restart service
sudo systemctl restart mosquitto

# Subscribe to a topic to verify
mosquitto_sub -h localhost -t "sensor/temperature" -v

For IoT devices like Raspberry Pi, you might configure a Python script to read a sensor and publish to MQTT:

import paho.mqtt.client as mqtt
import time

client = mqtt.Client()
client.connect("edge-gateway.local", 1883, 60)

while True:
    temperature = read_sensor()  # hypothetical function
    client.publish("sensor/temperature", temperature)
    time.sleep(10)

Verification can be done using mosquitto_sub on the edge gateway to see incoming messages. Network connectivity can be tested with ping and traceroute.

Edge computing often works alongside cloud computing, fog computing, and IoT. Fog computing is an intermediate layer between edge and cloud, often using local area network (LAN) infrastructure like routers and switches to provide compute, storage, and networking. The distinction: edge computing occurs on the device or gateway, while fog computing occurs on the LAN infrastructure. Both aim to reduce latency and bandwidth usage.

IoT devices are the primary data sources for edge computing. They can be sensors, actuators, wearables, or any internet-connected device. Common IoT communication protocols include: - Zigbee: Low-power, mesh network, used in home automation. - Z-Wave: Similar to Zigbee but primarily for home automation. - Bluetooth Low Energy (BLE): Short-range, low-power, used for wearables and beacons. - LoRaWAN: Long-range, low-power, for wide-area IoT. - NFC: Near-field communication for short-range, tap-to-connect scenarios.

Edge computing also integrates with cloud services like AWS IoT Greengrass, Azure IoT Edge, and Google Cloud IoT Edge, which provide software to manage edge devices and synchronize with the cloud.

For the 220-1101 exam, focus on:

Trap: Some candidates confuse edge computing with fog computing. Remember: edge is at the device/gateway level; fog is at the LAN infrastructure level (routers, switches). Another trap: assuming all IoT devices use Wi-Fi; many use low-power protocols like Zigbee or BLE.

A large automotive manufacturer deploys hundreds of vibration sensors on assembly line motors. Each sensor samples at 1 kHz and sends raw data to an edge gateway via BLE. The gateway runs a Fast Fourier Transform (FFT) to detect frequency signatures indicative of bearing wear. If a signature exceeds a threshold, the gateway sends an alert to the maintenance team and automatically adjusts the motor speed to prevent failure. Only the FFT results and alerts are sent to the cloud for long-term trend analysis. This reduces cloud data volume from 1 GB/hour to 10 MB/hour. Misconfiguration: If the threshold is set too low, false alarms flood the system; too high, and failures are missed. The edge gateway must be sized to handle the compute load—a Raspberry Pi would be insufficient; a Jetson Nano is often used.

A retail chain uses RFID tags on products and IoT readers at shelves. Each reader reports tag presence every second. An edge server in the store aggregates data and updates inventory counts locally. When a product is low, the edge server triggers a restock alert to store staff. Only daily inventory snapshots are sent to the cloud for corporate analytics. The edge server also handles local failover: if the internet connection is lost, the store continues to operate normally, and data syncs when connectivity returns. Common issue: RFID reader interference from metal shelves—solved by adjusting reader power and antenna placement.

A commercial building uses temperature, occupancy, and light sensors to control HVAC and lighting. An edge gateway runs a rule engine: if a room is unoccupied for 15 minutes, it sets the thermostat to setback temperature (e.g., 18°C) and turns off lights. The gateway uses BACnet or Modbus to communicate with the HVAC system. Cloud integration allows facility managers to view aggregated energy usage and adjust schedules. Performance consideration: The gateway must handle hundreds of devices with sub-second response. Common failure: misconfigured occupancy sensors causing rapid cycling of HVAC—solved by adding a hysteresis timer (e.g., 5-minute delay).