Build your First IoT Application with IBM Watson IoT

Post on 21-Jan-2018

92 views 5 download

Transcript of Build your First IoT Application with IBM Watson IoT

Building IoT Applications with IBM Watson IoT

Janakiram MSV@janakiramm

Tech Talk

IoT 2010

IoT 2017

The Big Picture of IoT

Devices

Device Registry

Data Ingestion

Public Cloud

Message Routing Policies

Storage&

Database

Stream Analytics

Batch Processing

Machine Learning

Business Intelligence

Remote Assets

Command Control

Admin & Management

User Expereince

Devices

• IP capable devices• Can directly connect to the platform

• Non-IP capable devices• Uses ZigBee, BLE, Z-Wave and PoE

• Legacy devices• Relies on SCADA – RTU, PLC, and

Telemetry

Gateway

• Gateways are two types• Field Gateway• Cloud Gateway

• Registers each device with unique identifiers

• Devices may have additional metadata• Each device is associated with

credentials for authentication• Gateway authorizes devices for

• Connecting• Publishing• Subscribing

• Devices can be whitelisted or blacklisted• Manages D2C and C2D messages• Exposes HTTP, MQTT, AMQP, and

WebSockets

Data Ingestion

• Ingestion endpoint are different from device management endpoint

• Enables high velocity, high throughput stream ingestion

• Enables one-way communication between devices and platform

Storage

• Uses a combination of object storage, NoSQL and relational databases

• Stores both raw data stream and processed datasets

• Acts as the source for queries that deliver operational insights

• Acts as the destination for processed dataset

• Exposes REST API for applications to access the data

Hot Path Analytics

• Transforms, processes, queries, and analyzes streaming data

• Connected to the ingestion layer as one of the subscribers

• Process each item in a stream at a time or using a sliding time window

• Modifies data format from one type to another

• Can invoke 3rd party web services for transforming data

Cold Path Analytics

• Delivers insights from data collected over a period of time

• Uses Map Reduce algorithm for processing data

• Leverages traditional Apache Hadoop

Applications

• End users and decision makers access the system through this layer

• Used for command, control, visualization and dashboards

• Implemented as mobile, wearables, web, and desktop apps

• Enables both retrieving the data and sending commands to devices

The Big Picture of IoT

Devices

Device Registry

Data Ingestion

Public Cloud

Message Routing Policies

Storage&

Database

Stream Analytics

Batch Processing

Machine Learning

Business Intelligence

Remote Assets

Command Control

Admin & Management

User Expereince

DemoBuilding an End-to-end IoT Application

Thank You!

Janakiram MSVjanakiram.com