MCA VI Semester
Internet of Things
TMC-403
Jaishankar Bhatt
Assistant Professor
Graphic Era Deemed to be University
Dehradun
Graphic Era Deemed to be Universsity, Dehradun
UNIT 4
Contents
• Data and its Analytics for IoT
• Introduction to Data Analytics for IoT
• Machine Learning, Big Data Analytics Tools and Technology
• Edge Streaming Analytics, Network Analytics.
• Securing IoT
• A Brief History of IoT Security
• Common Challenges in IoT Security
• How IT and IoT Security Practices and Systems Vary.
Introduction to Data Analytics for IoT
What is IoT?
• Internet of Things (IoT) refers to interconnected devices that collect and exchange data via the internet.
• Examples: Smart home devices, industrial sensors, connected vehicles.
Role of Data in IoT
• IoT generates massive amounts of real-time data from diverse sources (sensors, GPS, RFID, etc.).
• Data can be structured, semi-structured, or unstructured.
Data Analytics for IoT
• Goal: Extract actionable insights from IoT data to drive decisions and automation.
• Steps:
o Data Collection (sensors/devices)
o Data Transmission (networks/protocols)
o Data Processing and Storage (cloud/edge/fog)
o Data Analysis (real-time/historical)
Importance
• Enables predictive maintenance, operational efficiency, real-time monitoring, customer behavior analysis, etc.
Machine Learning for IoT
Why ML in IoT?
• IoT data is massive and complex — ML helps to uncover patterns and make predictions.
ML Techniques Used in IoT
• Supervised Learning (e.g., classification of sensor failures)
• Unsupervised Learning (e.g., clustering usage behavior)
• Reinforcement Learning (e.g., smart energy management)
Common Use Cases
• Predictive Maintenance: Predict failures using historical sensor data.
• Anomaly Detection: Detect abnormal behavior in connected systems.
• Smart Cities: Optimize traffic and utilities.
• Healthcare: Remote patient monitoring with anomaly alerts.
ML Pipeline for IoT
1. Data Collection
2. Preprocessing (cleaning, normalization)
3. Feature Engineering
4. Model Training
5. Deployment (often on edge/fog devices)
Big Data Analytics Tools and Technology
Key Characteristics of IoT Data (Big Data)
• Volume: Large amount of data
• Velocity: High-speed generation and processing
• Variety: Different formats (text, images, video, sensor data)
• Veracity: Uncertainty in data accuracy
• Value: Extracting useful insights
Tools & Technologies
Category Examples
Hadoop Distributed File System (HDFS), Amazon
Data Storage
S3, Apache Cassandra, MongoDB
Data Processing Apache Hadoop, Apache Spark, Flink, Storm
Stream Processing Apache Kafka, Apache Storm, Apache Flink
ML Libraries TensorFlow, Scikit-learn, PyTorch, MLlib (Spark)
Visualization Tableau, Power BI, Grafana
Platforms: AWS IoT, Azure IoT Hub, Google Cloud IoT — offer end-to-end support for analytics.
Edge Streaming Analytics
What is Edge Analytics?
• Analytics done near or at the data source, rather than in a centralized cloud.
Need for Edge Analytics
• Reduces latency, bandwidth usage, and ensures real-time responsiveness.
• Ideal for time-sensitive applications (e.g., autonomous vehicles, industrial safety).
Edge Analytics Workflow
1. Data captured by IoT sensors
2. Data processed by edge device/gateway
3. Processed data transmitted to cloud (if needed)
Technologies
• Edge devices: Raspberry Pi, NVIDIA Jetson, Intel NUC
• Frameworks: AWS Greengrass, Azure IoT Edge, EdgeX Foundry
Use Cases
• Smart manufacturing (detecting faults in real time)
• Retail (real-time customer behavior tracking)
• Remote monitoring (oil rigs, power stations)
Network Analytics in IoT
What is Network Analytics?
• Analyzing communication patterns, traffic loads, and performance across IoT networks.
Importance
• Ensure reliability, security, and performance of IoT data transmission.
• Detect network anomalies or failures early.
Key Metrics Analyzed
• Latency, bandwidth, packet loss, device connectivity, signal strength
Tools and Techniques
• SNMP-based tools, NetFlow, Wireshark, network telemetry systems
• AI/ML for:
o Predictive network performance
o Adaptive routing
o Anomaly and intrusion detection
Use Cases
• Smart Grids (ensuring stable communication between sensors and control
centers)
• Industrial IoT (monitoring connectivity between PLCs and cloud)
• Healthcare IoT (ensuring uninterrupted patient data streaming)
A Brief History of IoT Security
Early Days of IoT
• IoT started as simple embedded systems for automation and monitoring (e.g., thermostats, RFID, wearable tech).
• Security was not a priority — focus was on functionality, cost, and battery life.
Evolution of IoT
• As connectivity increased (Wi-Fi, Bluetooth, 4G/5G), IoT devices became smarter and more widespread:
• Smart homes, connected cars, health trackers, industrial sensors.
• Security risks emerged: devices connected to the internet without protection.
Rise of Threats
• Mirai Botnet (2016): Major attack that used insecure IoT devices to launch massive DDoS attacks.
• BrickerBot, VPNFilter, and other malware exploited weak/default credentials and outdated firmware.
• Governments began regulating IoT security (e.g., UK’s IoT cybersecurity law, California’s SB-327).
Current Focus
• Shift toward “Security by Design”: embedding security into devices from the start.
• Use of secure boot, hardware root of trust, encryption, and firmware updates.
Common Challenges in IoT Security
1. Weak Authentication
• Many IoT devices still use default or hardcoded passwords.
• Lack of multi-factor authentication (MFA).
2. Insecure Communication
• Data often transmitted unencrypted (e.g., HTTP instead of HTTPS).
• Vulnerable to man-in-the-middle (MitM) attacks.
3. Lack of Updates
• IoT devices may not support over-the-air (OTA) firmware updates.
• Users are often unaware or unable to update devices, leaving vulnerabilities
unpatched.
4. Device Constraints
• Limited processing power, memory, and storage.
• Difficult to implement strong encryption or complex security protocols.
5. Supply Chain Risks
•Firmware or components may have malware/backdoors before deployment.
•Lack of transparency in manufacturing and sourcing.
6. Scalability
•Large-scale deployments (e.g., smart cities) make it hard to manage and monitor every device.
•Risk of botnets, DDoS, and automated attacks increases with scale.
7. Privacy Concerns
•IoT devices collect sensitive personal data (location, voice, video).
•Weak protection can lead to data breaches, identity theft, and surveillance.
How IT and IoT Security Practices and Systems Vary
Features IT security IoT Security
Primary Goal Data confidentiality, integrity, availability Device reliability, data privacy, uptime
System Complexity Managed computers and servers Massive number of diverse smart devices
Update Mechanism Regular software patches & updates Often lacks OTA updates or user support
Authentication Strong (passwords, MFA, SSO, etc.) Often weak or default authentication
Security Tools Firewalls, antivirus, endpoint protection Lightweight or absent due to constraints
Network Design Centralized Distributed and dynamic
Lifecycle 3–5 years 5–15 years, often unsupported after sale
User Interaction Tech-savvy admins Non-technical end users (home or field)
Patch Management Automated or manual patching Often no patching mechanisms
Key Implications
• IoT security requires lightweight, scalable, and automated protection mechanisms.
• IT policies must be adapted to fit IoT’s hardware and connectivity limitations.
• Secure IoT design must consider long device lifecycles, physical access, and non-technical users.
Summery of the Unit 4
Area Focus
Data Analytics for IoT Extract insights from device-generated data
Machine Learning Predictive models for sensor behaviour, maintenance, and detection
Big Data Tools Handle the scale, speed, and complexity of IoT data
Edge Streaming Analytics Process data close to where it’s generated
Network Analytics Ensure stable, efficient, and secure data transmission in IoT
IoT Security
• IoT security has evolved from being an afterthought to a critical requirement due to real-world attacks.
• Unique challenges such as weak authentication, lack of updates, and constrained devices make IoT harder to
secure than traditional IT.
• Security practices need to be tailored for IoT environments, considering hardware limits and usability, while
following core principles of cybersecurity.
