Making sense of IoT, M2M and Big Data

Image Source: IEEE Spectrum

The IoT Architecture

Information Value

Loop

ApplicationsAnalyticsData AggregationNetworksSensors

M2M

Big Data

IoT

The IoT Equation

Sensors – The Eyes & Ears

Cost of a Sensor

µ-Processor Clock Speed

Key Issues: Power Constraints Energy

Harvest Data Transfer Event-driven Security Encryption,

Bandwidth Interoperability Std. vs Prop.

Small, Low Power, Low Memory Devices

Networks – The Lifeline

Home Area Network(Bluetooth, 6LoPAN, ZigBee, WiFi)

Neighborhood Area Network

(WiFi-Mesh, Cellular)

Wide Area Network(Fixed, Cellular WAN)

Key Issues: Coverage Range, Reliability Capacity Uplink-biased, Protocol

Overhead Security Authentication, Privacy Multiple Access Last Grasp, Sleep

Mode Addressing IPv6, Multicast/Broadcast Protocol s MQTT, COAP QoS Variability OFDM, DSCP

Too Many Devices, Too Many Types

Data Processing – The Brian

SQL/NoSQL Queries

Key Techniques: Distributed Storage (Hadoop HDFS) Parallel Processing (MAP-Reduce, Apache

Pig/Hive/HBase) In-memory Processing (Apache Spark)

Too Many, Too Much, Too FastDATA BIG DATA

Analytics – The Mind

Open-source Analytics Tools

Key Areas:Artificial Intelligence (AI)Machine LearningDeep LearningComplex Event Processing (CEP)Natural Language Processing (NPL)

Making Sense of Data

Applications – The Life of IoT

Case Study – Connected Vehicles

Autopilot Smart Charging

Solar Charging

Vehicle to Grid