MVP Summit 2013 Exhibit Session

1. SmartHome and
SmartFactory
Intelligent Systems
With .NET Micro Framework, Azure and MQTT
Lorenzo Maiorfi
Mirco Vanini

2.  Cloud Hosting Environment (Azure VMs)
 MQTT Network
 MQTT Broker by HiveMQ (www.hivemq.com)
 .NET MF embedded devices running as MQTT Network “nodes”
Our Reference
Application
Modules
 sensors/actuators working as publishers/subscribers of
data/commands
 MQTT-2-Z-Wave embedded bridge
 MQTT-2-ZigBee embedded bridge
 MQTT-2-ISM (2.4GHz Radio) embedded bridge
 MQTT industrial controller (ABB PLC)
 Node-red “engine”, acting as
 designer and runtime host for “flows”
 interface/protocol/API adapter among service endpoints
 simple business-logic modules hosting platform

3.  Application protocol over TCP layer
 Based on pub/sub messaging model
MQTT is the
de-facto
standard for
M2M / IoT
Applications
 Messages are made of two simple parts:
 Topic, a string typically expressed in hierarchical form
“/part1/part2/part3/…”
 Payload, an arbitrary content expressed as a byte-array
 3 QoS levels for message publishing
 QoS Zero (best effort)
 QoS One (broker confirms to publisher that message has been
delivered, but cannot confirm that same message wasn’t delivered
more than once)
 QoS Two (broker confirms to publisher that message has been
delivered, and just once)
 MQTT provides for both “retained” and “last will” messages

4.  Blazingly Fast
 Enterprise-grade security
HiveMQ
Broker
by DC2
 Clustering
 Portability
 Support
 100% MQTT compliant
 Websockets
 Scalability
 Modular
 Administration made easy

5.  Despite its powerful network model, real world scenarios often
require to extend such network to other communication systems
 A common requirement is radio coverage:
MQTT
Bridging
 ZigBee mesh networks for industrial environments
 Z-Wave mesh networks for smart home systems
 ISM (2.4 GHz) technology for cheap and fast data exchange
 Bridging MQTT with these technologies allows for:
 building hybrid networks whose nodes share same messaging model
 joining different MQTT networks with radio links
 greater overall reliability (by redundancy)

6.  “Mesh” networking model means that any node can extend
overall coverage to adjacent nodes, acting as a router
 Digi XBee modules are by far the most common chips used
worldwide
ZigBee
Networks
 Coverage is about 50-100m (1mW) for standard modules and up to
1.6km (50mW) for “Pro” modules
 They’re great to implement remote IO management (digital
inputs/outputs and analog inputs)…
 …but together with a host microcontroller board they’re great to
implement arbitrary message exchange too
 They can be configured and updated remotely
 Serial API exposed to host applications is simple but powerful
 We have extended XBee API client libraries originally coded by
Micheal Schwartz (mfsmarttoolkit.codeplex.com) in order to get
greater reliability and performance

7.  Licensed by Sigma Designs (formerly ZenSys), Z-Wave is
becoming the “de facto” standard for Smart Home systems
 Z-Wave is based on mesh network model (like ZigBee), but radio
modules typically require much less power to work than XBee (a
battery operated device can work for years)
Z-Wave
Networks
 There are plenty of Z-Wave enabled devices (power outlets, power
meters, sensors, thermostats, blinds motors drivers, switches, etc)
 We developed a managed Z-Wave Serial API Controller stack
targeted to .NET Micro Framework devices
 Our client lib provides support for most commons profiles (basic,
meter, multilevelsensor, multilevelswitch, wakeup, etc)
 Client API is async-only
 Works with any Z-Wave Serial API Controller device, connected
through a UART port to host application microcontroller

8.  Despite its great flexibility, a network infrastructure itself cannot
address real-world scenarios
 We need a way to design and run business logic flows
orchestrating message exchange among network nodes
Business Logic
Flows
 Node-red is an open project whose main purpose is that to act as
an engine that runs/hosts “flows” designed by mean of node-red
itself
 Despite being general, it fits in a natural way with MQTT
 Node components can be written in JavaScript
 It uses a visual design language using a SPA web IDE (hosted by
node-red itself)
 Many node components are built-in
 Despite it can also implement business logic of medium complexity,
node-red killer-application is no doubt “interface/API” adaptation
among “actors” in heterogeneous systems (in our reference
application we provide an example of both cases)

9.  Windows 8 Store App
 consumes a custom tcp socket “tunneling” in order to get and
publish MQTT messages even with no MQTT client lib support
Sample Client
Apps
 Windows Embedded Compact 2013 App
 uses µM2MQTT Library by Paolo Patierno in order to directly
interact with MQTT broker
 Windows Phone 8, iPhone Apps
 use a plain HTTP REST/JSON API
 kind of “longpolling” technique implemented by a node-red flow
changes messaging model from pub/sub to request/reply

10.  Mirco Vanini: Windows Store App
 Paolo Patierno: uM2MQTT Library & WEC 2013 Demo App
 Fabrizio Bernabei, Tiziano Cacioppolini: WP8 Demo App
Credits
 Daniele Galiotto: iPhone App
 Lorenzo Maiorfi: firmware, web broker spa web app, node-red
flows