hoc and Wireless Sensors Networks BIT 5th Annual World Congress Emerging InfoTech 2016 An IOT gateway Architecture using a MQTT bundle transport & LTP Convergence Layer 11/11/16 1 By Alex Kritikos
OASIS MQTT TC member Email: [email protected] Twitter: @krital LinkedIn: https://uk.linkedin.com/in/alexkritikos ALEX KRITIKOS About the speaker 2 Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks
hoc and Wireless Sensors Networks Earth & Space IOT research What if we combined them? • Link layer agnostic propagation (bundles) • MQTT encapsulation: payloads to local subscribers 5 11/11/16 • DTN encapsulation: payloads are routed progressively via intermediary nodes • Full MQTT & DTN stack on each node • Lower latency and higher throughput
= cost reduction • Software sharing = cost reduction • MQTT bridging • via DTN BP • Via Licklider Transmission Protocol (LTP) DTN to MQTT Value Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks
local fanout from bridging propagation 11 11/11/16 • Events are routed progressively via intermediary nodes • Results in lower latency and higher throughput MQTT ROUTING via DTN Thoughts Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks
control via LTP sessions • Multi stream (red, green) assured data propagation • Link layer agnostic 13 11/11/16 MQTT ROUTING via LTP Thoughts Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks
collaborate 14 11/11/16 Earth & Space IOT research What if we combined them? Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks
hoc and Wireless Sensors Networks References • Advances in Delay-tolerant Networks (DTNs) – Architecture and Enhanced Performance By Woodhead Publishing 2015 11/11/16 15 • Brenda Lyons, NASA NDBS 2016 • Appendix A: What is IOT? • Appendix B: What is DTN? • Appendix C: What is MQTT? • Further Research A: DTN routing via MQTT • Further Research B: High availability with MQTT • Further Research C: High availability with MQTT & DTN
- Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 16 11/11/16 • Read about • History and status • M2M Comparison • Market & estimates • Example architectures
the telecom era • M2M: proprietary, closed, private machine communication (e.g. CallerID) • IOT is much more than machine communication Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 17 11/11/16
a larger network, usually via a gateway, are cloud enabled, can interop, enables device physical reach (e.g. Samsung SmartThings) • IOT – M2M common: machine comms Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 18 11/11/16
• Today: • 6.4-9 billion Connected devices !This excludes phones, tablets, PCs! Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 19 11/11/16
COAP, AMQP, Websocket, Node • Enormous telco traffic challenge • Global M2M traffic:1 to 6.3 Exabytes / month by 2020 SOURCE • Real time streaming analytics • Security nightmare • 1TB DDOS by hacked CCTVs • PCs, PVRs, routers, smart fridge • MCUs + PKI = slow • Privacy concerns Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 20 11/11/16
Future Development of Ad hoc and Wireless Sensors Networks 22 11/11/16 Why cant all things use WiFi? • If mains powered, sometimes they do • WiFi may not reach your boiler room • A lot of battery powered – low energy sensors • WiFi lacks ad-hoc network features (e.g. mesh) Things tend to be in their own sensor/field net • Based on low power RF (e.g. zigbee, zwave) • Needs a gateway to bridge to local IP network
/ WiFi HTTP MQTT AMQP Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 23 11/11/16 Users interact with things via a device …
RF ZigBee ZWave BLE Ethernet / WiFi HTTP MQTT AMQP Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 24 11/11/16 Things interact with environment via gateway…
RF ZigBee ZWave BLE Ethernet / WiFi HTTP MQTT AMQP Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 26 11/11/16
on the Future Development of Ad hoc and Wireless Sensors Networks 27 11/11/16 • Read about • history and status • Protocol overview • TCP/IP Comparison • Terrestrial uses
Development of Ad hoc and Wireless Sensors Networks 28 11/11/16 initiative to develop an “Interplanetary Internet” 1998 NASA starts Delay / Disruption Tolerant Networking 1998 NASA leads worldwide DTN development 2005 IOAG Recommends Solar System Internetwork (SSI) 2010 IOAG Suggests DTN a core service of SSI 2010 DTN is being standardized by IETF DTN is being standardized by CCSDS
Internet capabilities • Works • in challenged communication environments • over unidirectional/asymmetric links • Tolerates • frequent disruptions • long delays / high error rates • Provides • assured delivery of data • automatic store-and-forward • Can run over IP or as a full Internetworking protocol Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 30 11/11/16 Overview
Can be used to propagate data not needed in real time (e.g. wearables, home automation etc) • Introduces some node storage requirements • Use cases • sea-going vessels • outdoor explorers • wildlife research • vehicular networks • disaster area networks Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 35 11/11/16 Terrestrial use
on the Future Development of Ad hoc and Wireless Sensors Networks 36 11/11/16 • Read about • History and status • Protocol overview • Role in IOT architectures
of Ad hoc and Wireless Sensors Networks 37 11/11/16 IBM & Eurotech develop 1990s IBM & Eurotech Open Source 2012 Software AG implements 2013 OASIS MQTT TC Created 2013 Many others join MQTT becomes OASIS standard 2014 MQTT becomes ISO/IEC 20922 2016
lightweight publish- subscribe binary messaging protocol • Is easy to implement (15 protocol commands) • has a fixed 2 byte over-the-wire overhead header • Has 256MB max payload • runs on top of TCP/TLS/WS/WSS stacks Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 38 11/11/16
levels: • QOS 0: At most once (Fire and forget) • QOS 1: At least once (ACKed delivery) • QOS 2: Exactly once (ACKed / No duplicates) Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 39 11/11/16
http://ow.ly/wHm2e • The Eclipse Paho project offers some great MQTT clients in minimal C, regular C, C++, Java, Javascript, Python, Lua, Go, Objective- C: http://ow.ly/wHl8c Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 40 11/11/16
Data sharing importance: • Vehicular networks • Disaster area networks • Delegates shared custodian to MQTT broker • Can use various MQTT QOS Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 47 11/11/16
PUB STREAM 1 MQTT PUB STREAM 2 MQTT PUB STREAM 3 Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 54 11/11/16
assured data propagation • Decouples local fanout from slave propagation • Bundles continue to flow while broker offline • Allows fast cluster recovery Alex Kritikos - Thoughts on the Future Development of Ad hoc and Wireless Sensors Networks 60 11/11/16