MTN

MTN is now called Blue Horizon.

MTN: analysis at the edge

This talk live: https://www.youtube.com/watch?v=_kTajbcAd9E

Project MTN is an IBM exploration of decentralised, autonomous edge computing.

A global network of raspberry pis is assembled to demonstrate blockchain-based negotiation between devices in the IoT, offering and contracting data streams to perform analysis close to the edge.

Take part and learn more: http://bluehorizon.network

Transcript

1. © 2015 IBM Corporation | 1 MTN Devcon 1 London 2015 http://tremor.hovitos.engineering slides version 1.0.2
2. © 2015 IBM Corporation | 2 MTN Multiplying Things Needlessly
3. © 2015 IBM Corporation | 3 • Mac Devine, VP Cloud Innovation, 
 Executive Owner
 • Egan Ford, Distinguished Engineer
 Project Architect
 • Robin Lindenmeyer
 Program Manager
 • 12 engineers Team MTN
4. © 2015 IBM Corporation | 4 Speaker IBM Systems Senior programmer Blockchain Group ! ! ! ! hdiedrich@de.ibm.com twitter @hdiedrich Distributed Databases High Velocity Scalability ! Insurance Payment Games Henning Diedrich
5. © 2015 IBM Corporation | 5 MTN: Multiplying Things Needlessly Autonomous Processing,
 Communication, 
 and Coordination 
 between the Edge and the Cloud
6. © 2015 IBM Corporation | We are at the threshold
   of a massive explosion in the history of computing. 2020 30 billion 2003 500 million 2009 2,5 billion 1975 10,000 2050 > 100 billion 2014 10 billion 6 1950 5000
7. © 2015 IBM Corporation | 7 The first wave of the IoT is already here. It’s the Internet of billions of things. Proportion of Smart TVs being used for Internet Viewing Continuous Monitoring of Jet Engines Automated Smart Meters Proportion of Heavy Industrial Equipment Networked 30% But demand has been slow to take off in many other areas We have seen visible successes in high value applications 10%
8. © 2015 IBM Corporation | 8 As  more  and  more  devices  around  us   become  connected  and  intelligent,   many  physical  products  as  we  know   them  will  be  transformed  into  digital   experiences.    
 
 Many  machines-­‐human  interactions   will  be  replaced  by  machine-­‐machine   (peer-­‐to-­‐peer  interactions).   MTN Motivation
9. © 2015 IBM Corporation | 9 Power in the IoT will shift from the center to the edge. ! As devices compete and trade autonomously in real-time, they
 will create liquid markets out of
 the digital marketplaces. ! Connectivity and intelligence will be a means to better products and experiences, not an end. A Device Democracy is emerging.
10. © 2015 IBM Corporation | p As devices become empowered to autonomously
    manage their roles and behavior, a democracy of devices will emerge. The blockchain, a universal digital ledger, facilitates transaction processing between devices. 10 Registration of new device Authentication of remote users Contract to barter power with other appliances Checklist for automobile safety
11. © 2015 IBM Corporation | IoT solutions that work for billions of devices
    won’t necessarily scale to hundreds of billions of devices. 11 Lack of Not future-proof Lack of privacy Internet of Things Brok en business High cost m odels functional value
12. © 2015 IBM Corporation | Connecting hundreds of billions of devices
    is going to be expensive. 12 Cost of centralized infrastructure and server farms. Cost of supporting and maintaining hundreds of billions of devices. Mismatch between cost and revenues. Lack of Not future-proof Lack of privacy Internet of Things Brok en business High cost m odels functional value
13. © 2015 IBM Corporation | The Internet is not a place of trust. 13 No trusted partners 
 on the internet. Centralized authorities become single points of failure. Trust can be very hard to engineer, impossible to guarantee. Lack of Not future-proof Lack of privacy Internet of Things Brok en business High cost m odels functional value
14. © 2015 IBM Corporation | Smart devices need to be managed
    for the long haul. 14 Average life of IoT devices will be decades. Average life of smart device manufacturers 
 5-10 years. Cost of maintaining smart devices can outlast product and manufacturer obsolescence. Lack of Not future-proof Lack of privacy Internet of Things Brok en business High cost m odels functional value
15. © 2015 IBM Corporation | Simply enabling connectivity
    doesn’t necessarily make a product better. 15 Value proposition of many smart products is low. Many smart features are not worth the extra cost and complexity to users. Successes in the market have improved the core functionality and user experience. Lack of Not future-proof Lack of privacy Internet of Things Brok en business High cost m odels functional value
16. © 2015 IBM Corporation | p Profits remain elusive
 as companies continue to apply broken business models. 16 Unrealistic expectations from revenue from software updates and apps. ! Unrealistic expectations of data and analytics. ! Unrealistic expectations of ecosystem opportunities and energy efficiency gains. Lack of Not future-proof Lack of privacy Internet of Things Brok en business High cost m odels functional value
17. © 2015 IBM Corporation | Motivation
    Open access IoT networks, distributed cloud 17 • It  is  our  view  that  IoT  is  quickly  moving  to  a  data  and   machine  learning  driven  future  where  simple  convenience   (IoT  today)  will  be  replaced  with  a  transformative  experience   where  IoT  is  better  integrated  with  all  devices  and  our  lives.   • IoT  devices  should  be  able  to  communicate  and  coordinate   in  a  fully  autonomous  manner  while  being  constrained  by   policy,  policy  that  is  user  defined  and  tuned  by  machine   learning.   • Smart  Contracts  for  Device  Coordination   • Firmware  Updates   • Pub/Sub  Models
18. © 2015 IBM Corporation | Consumers buy solutions
    rather than networks or business models. 18 Make it better Networking and remote usage 
 will become secondary features. ! Tomorrow’s smart appliances will create value by applying intelligence and network connectivity to improve the CORE value proposition of the device. Better cooking Increased safety Improved reliability Greater fitness Less mess
19. © 2015 IBM Corporation | To be secure and scalable
    the IoT will evolve to be more open and decentralized
 Design for Peer-to-Peer systems Design for Trustless communication Design for Decentralized Autonomy Open access IoT networks, distributed cloud Design principles of ADEPT architecture (Autonomous Decentralized Peer-to-Peer Telemetry) 19 Before 2005
 Today
 2025 and beyond
 Closed and centralized IoT networks Open access IoT networks, centralized cloud Open access IoT networks, distributed cloud
20. © 2015 IBM Corporation | A truly decentralized IoT
    will require 4 foundational functions Open access IoT networks, distributed cloud 20 • Trustless peer-to-peer messaging • Distributed file-sharing • Autonomous device coordination • Identity
21. © 2015 IBM Corporation | 21 Internet of Things 9bn devices today 100bn tomorrow
22. © 2015 IBM Corporation | 22 MTN Mission: Analytics at the Edge The  Edge  is  where  the  data  is  generated,  however
 
 ‘Analytics  at  the  Edge’   is  really   ‘as  close  to  the  Edge  as  possible  including  the  Edge’.
23. © 2015 IBM Corporation | 23 Massive Untapped Edge Data Need Analysis
 Data at the Edge 90% Volume of data created at the edge of IoT that is never captured, analyzed or acted upon. 60% Amount of that data that loses its value within milliseconds of being generated.
24. © 2015 IBM Corporation | 24 • Real-­‐time  processing  at  the  Edge   • Open,  multiparty,  autonomous  IoT  ecosystems   • The  behavior  of  decentralized  systems   • What  technical,  scalability,  stability,  security,  privacy,   economic,  and  social  problems  and  opportunities  can  be   identified?   • Containers  in  a  hostile  environment   • Better  quality  of  Experience? Exploration
25. © 2015 IBM Corporation | 25 Analysing Data at the Edge ! • Software Defined Radio (SDR) • 50 megabits/second of data at the edge • volume of data cannot be sent to the cloud
 or real-time analysis => • policy or algorithm-based analysis and/or filtering
 has to be done at the edge. MTN Proof of Concept
26. © 2015 IBM Corporation | 26 SDR
27. © 2015 IBM Corporation | 27 SDR
28. © 2015 IBM Corporation | 28 Terabytes of Data to Experiment With
29. © 2015 IBM Corporation | 29 SDR 12H Wide Spectrum Scan FM LTE Aircraft  Radio  Navigation Satellite   Mobile   Comm Smart  Meters  (Power,  Water,  Gas) ADS-­‐B
30. © 2015 IBM Corporation | 30 SDR Water and Power Meter Output {Time:2015-09-11T21:07:18.225 SCM:{ID:30969548 Type: 7 Tamper:{Phy:02 Enc:01} Consumption: 8605529 CRC:0xBA97}}! {Time:2015-09-11T21:07:23.657 SCM:{ID:26353424 Type:13 Tamper:{Phy:00 Enc:02} Consumption: 2270 CRC:0x5DAF}}! {Time:2015-09-11T21:07:35.782 SCM:{ID:30971260 Type: 7 Tamper:{Phy:01 Enc:00} Consumption: 1300037 CRC:0x2526}}! {Time:2015-09-11T21:07:37.678 SCM:{ID:30971322 Type: 7 Tamper:{Phy:02 Enc:00} Consumption: 821676 CRC:0xBE72}}! {Time:2015-09-11T21:07:40.177 SCM:{ID:56438625 Type: 7 Tamper:{Phy:02 Enc:02} Consumption: 5542565 CRC:0xD334}}! {Time:2015-09-11T21:07:41.595 SCM:{ID:56443013 Type: 7 Tamper:{Phy:02 Enc:01} Consumption: 781466 CRC:0x811B}}! {Time:2015-09-11T21:07:42.354 SCM:{ID:26353424 Type:13 Tamper:{Phy:00 Enc:02} Consumption: 2270 CRC:0x5DAF}}! {Time:2015-09-11T21:07:42.979 SCM:{ID:30975357 Type: 7 Tamper:{Phy:02 Enc:01} Consumption: 3769334 CRC:0x1886}}! {Time:2015-09-11T21:07:45.688 SCM:{ID:31032332 Type:13 Tamper:{Phy:00 Enc:01} Consumption: 216 CRC:0xD4DF}}" {Time:2015-09-11T21:07:49.125 SCM:{ID:26353424 Type:13 Tamper:{Phy:00 Enc:02} Consumption: 2270 CRC:0x5DAF}}! {Time:2015-09-11T21:07:52.853 SCM:{ID:31032332 Type:13 Tamper:{Phy:00 Enc:01} Consumption: 216 CRC:0xD4DF}}" {Time:2015-09-11T21:07:58.369 SCM:{ID:26353424 Type:13 Tamper:{Phy:00 Enc:02} Consumption: 2270 CRC:0x5DAF}}
31. © 2015 IBM Corporation | 31 SDR Flight Data
32. © 2015 IBM Corporation | 32 ADS-B
33. © 2015 IBM Corporation | 33 • Device  –  a  Raspberry  Pi  capable  of   running  a  docker  container   • Device  Contract  –  an  ethereum   smart  contract  that  represents  the   device  in  the  blockchain   • Token  Bank  Contract  –  a  smart   contact  that  holds  token  balances   for  ecosystem  participants   • Device  Registry  –  a  smart  contract   that  knows  about  all  device   contracts  and  characteristics  of  the   devices  represented  by  those   contracts   • Glensung  –  a  producer  of  docker   containers  and  a  consumer  of   docker  capable  devices  for  sale Personas
34. © 2015 IBM Corporation | 34 PoC IoT Infrastructure Big Picture
35. © 2015 IBM Corporation | 35 Terabytes of Data to Experiment With
36. © 2015 IBM Corporation | 36 Terabytes of Data to Experiment With
37. © 2015 IBM Corporation | 37 Terabytes of Data to Experiment With
38. © 2015 IBM Corporation | 38
39. © 2015 IBM Corporation | 39 Terabytes of Data to Experiment With
40. © 2015 IBM Corporation | 40 Initialisation – Bootstrapping Token  Bank   (contract) Device  Registry   (contract) RPi  Device   (contract) Ethereum Glensung Install  device  contract(ethereum  account) Register  device(contract  address,  device  characteristics) contract  address Mint  some  tokens Device  is  available  to  run  containers RPi  Device Associate  Token  Bank Check  my  token  balance token  balance write read
41. © 2015 IBM Corporation | 41 Running Containers Create  escrow  proposal(tokens) Find  a  device Make  a  proposal  to  run  a  container(container  pointer,  number  of  tokens  for  payment) Token  Bank   (contract) Device  Registry   (contract) RPi  Device   (contract) Ethereum Glensung RPi  Device Poll  for  proposal Proposal contract  address Download  container Vote  to  accept  proposal Poll  for  acceptance Accepted  by  device Vote  to  accept  proposal Poll  for  acceptance Proposer  accepted Proposal  complete Tokens  transfer  to  device  owner Device  ready Execute  container Phone  home Execution  complete write read
42. © 2015 IBM Corporation | Risks & Opportunities
    Open access IoT networks, distributed cloud 42 • Spot  market   • Control  access  to  sensors   • 3rd  party  access  to  raw  sensors   • How  do  vendors  want  to  access,  processes,  analyze,  filter,   sensor  and  how  do  they  want  to  stream   • Arbitrary  code
43. © 2015 IBM Corporation | MTN: Next Steps
    Open access IoT networks, distributed cloud 43 • Thin  Client  for  Embedded  Systems   • Exchange/Directory   • Internal  Socialization  Work   • Get  20,000  volunteers  to  help  test  scalability   • http://tremor.hovitos.engineering  
44. © 2015 IBM Corporation | Thanks to the Go Team!
    Open access IoT networks, distributed cloud 44 Jeff  Wilcke   Felix  Lange   Gustav  Simonsson   Viktor  Trón
45. © 2015 IBM Corporation | Thank You!
    Open access IoT networks, distributed cloud 45 Take  part  in  the  MTN  experiment:   http://tremor.hovitos.engineering     ! This  talk  live:   https://www.youtube.com/watch?v=_kTajbcAd9E