cockpit for big systems and big iot systems leveraging ibm bluemix and watson
TRANSCRIPT
Local Touch - Global Reach.
Intelligent System for Precision FarmingA partnership project between Drotek & SogetiLeveraging IBM Bluemix and Watson Technologies
World of Watson – Octrober 24-27 – Las Vegas
SOGETI France - 2016
IBM World of Watson 2016 Conference
Session ID: IOT-2052
Partnership Project – IBM World of Watson 2016 Conference
2 The Speakers
Patrick started programming on a Casio FX-602P and after on HP-41C, a long time ago! After more than a quarter century working on machines for people and large organizations, as a developer, as an architect, as a project manager or consultant he is still fascinated by the beauty and by the monstrous and elusive nature of complex systems, and in particular socio-technical systems. In 2012, as an innovation leader at Sogeti and members of the SogetiLabs, he launched five R&D programs using big data, software system testing, multi-agent systems, cloud computing, and project management as fields of experiments. His conviction is that instead of undergoing complex systems we must try to understand them, better control them, innovate, and do business with that.
Kevin is Drotek’s CTO & Associate. Technology aficionado, late night explorer, language learning addict, he is a real Swiss army knife for the start up. He loves it when someone comes up with a new « out of the blue » idea (if he doesn’t himself) and he has to design a prototype for it from scratch. Fast. Innovate. Iterate. That’s his motto. Skydiver, pilot, there seems to be no limit for him. He strongly believes that without risk there can be no success. Most important thing about his personality is that he is relentlessly eager to learn. He has proved his ability to solve complex problems in record time, as an intern he designed, developed and started to sell a disruptive product in less than 6 months.
Partnership Project – IBM World of Watson 2016 Conference
3 Once upon a time …
At different scale …
But still complex systems !
With unpredictable behavior and
evolution!
Complexity !
Partnership Project – IBM World of Watson 2016 Conference
4
ConstituentEcosystem
Why Complexity? Which issues?
System(of interest) Constituent
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
applies toapplies to
applies toapplies to
aggregation of
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
Sy stem Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
Sy stem Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
Behavior
Evolution
Plurality
Variety
Hierarchy
Feedback
Speed of changes
Emergence of Behavior
FACTS
Representation& Modelization
Calculability:Optimization& Prediction
ISSUES
Partnership Project – IBM World of Watson 2016 Conference
5
ConstituentEcosystem
System(of interest)
Constituent
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
applies toapplies to
applies toapplies to
aggregation of
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
Behavior
Evolution
The Agent and Multi-Agent Based System Paradigm
Associate System-Agent with Software-Agent
Wherever and Whenever possible Models based Agents
Partnership Project – IBM World of Watson 2016 Conference
6 Control and Command of Complex Systems
ConstituentEcosystem
System(of interest)
Constituent
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
applies toapplies to
applies toapplies to
aggregation of
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
System Constituant
Behavior
Evolution
Behavior
Evolution
Control
from
Sensors
Command
to
Actuators
Intelligence
Partnership Project – IBM World of Watson 2016 Conference
7
Historically, CBS started with the need to control-command and manage Very Large Systems of Systems (~100K machines) in the IT area, with a much larger number of assets (~20 managed assets per machine) and Ultra Large Scale System as final target (more than 1 million machines ~20 million assets) . Our concerns were:
• Efficient functioning with the same quality at different scales• Inherent heterogeneity of management of assets and diversity of modes to communicate with each others• Adjustment to the fine granularity of information in order to follow in real time the business needs• Orchestration control of the systems of systems
We built a generic infrastructure system, adaptable to business needs, and to use cases. CBS is OK with VLSS, but we are still running at the ULSS.
We need more scalability, elasticity and fast application development for experiments. We needed something like Bluemix!
Cockpit for Big System (CBS): The First Step
Partnership Project – IBM World of Watson 2016 Conference
8
• Currently , each instance of CBS can manage ~ 30 000 machines (more than half million assets managed)
• Monitoring information are retrieved each 15 minutes, then consolidated in a daily inventory 3Go of data consolidated daily 4To of data on a 18 months window
• + 250 000 assets inventoried in less than 5 minutes on a common LAN/WAN
• ~ 200 unique inventory keys, allowing to repertory, e.g., more than 700 applications, several OS, CPU, GPU, network interface controller, … Based on a generic inventory keys system, new sensors or actuators in the new version of CBS called CBIoTS can be added in some minutes.
• + 100K SLOC of Javascript, Perl, C, sh
• + 10 years of maturity of the agent software, which memory footprint is ~ 500Ko on Linux and ~ 1Mo on Windows
• + 20 environments / OS handled (Unix, Linux, Windows, Android, Raspberry PI, Yocto) . Soon on IOS and OSX
• All communications (agent-agent, agent-platform) use key authentication (AES) and communications encrypted (RSA)
Cockpit for Big System: Some Metrics
Partnership Project – IBM World of Watson 2016 Conference
9 Cockpit for Big Systems: Architecture Overview
CBS: Web Server Portal• Apache• Js/HTML• Widgets• Perl
CBS: Databases Cluster
• SQL• NoSQL• Graph
CBS: Consolidation Servers
• Consolidation• Operations automation
CBS: Data Collection Servers
• Data Collection• Data Integration• Listener
External Repositories
• LDAP• CMDB• …
CBS: Secure Gateway
• End-to-end connection• Encrypted tunnel• Encrypted communication
CBS real time compatible Agent• UNIX (Oracle, HP , IRIX, …)• LINUX (Red Hat, Cent OS,
Debian, Raspbian, …)• MS WINDOWS (all)• Yocto (Intel Edison)• Andoid• Soon on IoS and OSX
IBM Bluemix
• Docker containers• DB Server• Services
Partnership Project – IBM World of Watson 2016 Conference
10
The IoT Experiments: reach the ULSSand take other kind of socio-technical SoS up
Cockpit for Big IoT Systems: The New Stage of CBS
• Instantiation of the CBS platform for very big IoT systems
• Processing of Big Data• Retrieve of data and function determination given
the sensors• Action on the environment using actuators• A sensor (or actuator) = a function
Functional composition and triggered actions
CBIoTS Engine
CBS
Agen
t
Sens
ors/
Actu
ator
s up
to 2
0/un
it
Devi
ce w
ith U
nix,
Li
nux,
Win
dow
s or
An
droi
d O
S
CBS Agent
Sensors/Actuators up to 20/unit
Device with U
nix, Linux, W
indows or
Android OS
CBS Agent
Sensors/Actuators up to 20/unit
Device with Unix, Linux, Windows or
Android OS
CBS Agent
Sensors/Actuators
up to 20/unitDevice with Unix,
Linux, Windows or
Android OS
CBS AgentSensors/Actuators
up to 20/unit
Device with Unix,
Linux, Windows or
Android OS
CBS Agent
Senso
rs/Actu
ators
up to 20/unit Dev
ice w
ith U
nix,
Linux,
Windows o
r
Android OS
CBS Age
ntSe
nsors/
Actuato
rs
up to
20/unit
Device w
ith U
nix,
Linux,
Windows o
r
Android OS
EnvironmentEnvironment
Environment
Environment
Envir
onmen
t
Envi
ronm
ent
Envir
onmen
t
Partnership Project – IBM World of Watson 2016 Conference
11
Business ApplicationBusiness
Application
Cockpit for Big IoT Systems: Architecture Overview
Up to 30 000 units
encrypted
encrypted encrypted
encrypted
CBIoTS Platform
Sensors or Actuators(up to 20/unit)
Device OS
CBS Agent
IBM Watson
Business Application
Data (cloudified (IBM Bluemix) or on-premise)
High Data Streaming on IBM Bluemix
CBIoTS Engine on IBM Bluemix
Feedback Loop Feedback Loop
Feedback Loop
Up to half million sensors/ actuators under control/ command per CBioTS Engine instance
Device OS
CBS Agent
The IoT Experiments: reach the ULSS and take other kind of socio-technical SoS up
Use of Watson as a Global Feedback Loop
Use CBIoTS on Bluemix & Watson as an Infrastructure to quickly develop intelligent IoT business applications
Partnership Project – IBM World of Watson 2016 Conference
12
• Very Large IT Systems monitoring (done and running on-premise)• DevOps (done and running on Bluemix , see NADO project)• Smart Farming (in progress, see partnership with Drotek)• Smart Cities (ongoing prospection)• Smart Villages (in progress, some services are in partnership with Droteck)• Home automation• Transport (ongoing prospection)• Better understanding of System of Systems and Complex Systems (Current links with
research laboratories in Toulouse)• …
CBS/CBIoTS: Use Cases
Partnership Project – IBM World of Watson 2016 Conference
13 CBS/CBIoTS: Partnership with IBM
Special thanks to the IBM Architect Team in India• Vishal A CHAREGAONKAR• Naveen K PRATHAPANENI• Prashant S NAIK• Shiva KUMAR H R
Which are helping us in the porting of the CBS/CBIoTS platform on Bluemix
Many thanks to all the people involved in our partnership with IBM• Christian CHATEAUVIEUX (IBM France)• Claude BRUSSOLO (IBM France)• Olivier LISE (SOGETI)• Philippe DURANTE (SOGETI)• Jacques MEZHRAHID (SOGETI)• Mark RIGGS (Capgemini)• And all the others not mentioned here
Partnership Project – IBM World of Watson 2016 Conference
14
The information contained n this presentation s proprietary.© 2015 Sogeti. All rights reserved.
www.fr.sogeti.com
About SogetiSogeti is a leading provider of technology and software testing, specializing in Application, Infrastructure and Engineering Services. Sogeti offers cutting-edge solutions around Testing, Business Intelligence & Analytics, Mobile, Cloud and Cyber Security, combining world class methodologies and its global delivery model, Rightshore®. Sogeti brings together more than 20,000 professionals in 15 countries and has a strong local presence in over 100 locations in Europe, USA and India. Sogeti is a wholly-owned subsidiary of Cap Gemini S.A., listed on the Paris Stock Exchange.
TMap®, TMap NEXT®, TPI® and TPI NEXT® are registered trademarks of Sogeti, part of the Capgemini Group.
Thank you for your attention!
Now, lets talk about Precision Farming with Kevin from Drotek…
Partnership Project – IBM World of Watson 2016 Conference
15 Planet Earth in 2050
+30%
+70%
+5%
+0.7%
People
Food need
Arable land
YieldOUR WORLD
IN2050
Partnership Project – IBM World of Watson 2016 Conference
16 Main Goals
COMPETITIVENESSECONOMICS
YIELD
ECOLOGY
SOCIAL
Partnership Project – IBM World of Watson 2016 Conference
17 Farmers overwhelmed by data
Partnership Project – IBM World of Watson 2016 Conference
18 ISPF: The Ecosystem
Partnership Project – IBM World of Watson 2016 Conference
19
IBM BluemixMobile Devices, Sensors & Actuators
Fixed SensorsISPF
Portal Clou
d or
On-
prem
ise D
ata
Drotek embedded system &
CBIoTS smart agents
A fully distributed Architecture based on Embedded Systems, Multi-
Agent and Cloud Technologies
TCP, LoRa, Sigfox
ISPF: Architecture overview
Partnership Project – IBM World of Watson 2016 Conference
20 ISPF in Real Life
Partnership Project – IBM World of Watson 2016 Conference
21 ISPF Measurable Metrics
10%15%
99%
5%
Fuel economySavings on seeds, fertilizerand chemicalsDecrease on fertilizer cost
Yield increase
Partnership Project – IBM World of Watson 2016 Conference
22 ISPF: Main script
From Sensors (S) to Actuators (A)
+
+
CBIoTSWATSON Network
CBIoTS Network
S
A
A
Network CBIoTS WATSON+1
2
3
1Data collection 2Automation with: • Context awareness• Smart behavior
3Analytics & Forecasting for • Context awareness• Smart content
Partnership Project – IBM World of Watson 2016 Conference
23 ISPF: Functional Overview
Animation
Partnership Project – IBM World of Watson 2016 Conference
24 Intelligent System for Precision Farming
Partnership Project – IBM World of Watson 2016 Conference
25 Thank you
Thank youFor more details contact:
Patrick MarquetPractice Leader Enterprise Solutions Consulting - Sogeti France
Partnership Project – IBM World of Watson 2016 Conference
26
The information contained n this presentation s proprietary.© 2015 Sogeti. All rights reserved.
Thank you for your attention !
drotek.com
About DROTEKDrotek is a French tech start-up made of technology aficionados and late-night explorers. We think, design and produce as if we still were in our garage. Fast. Innovate. Iterate. This is our motto. We love it when someone comes up with a new "out of the blue" idea and we have to design a fully-working prototype completely from scratch. We strongly believe that people should pay for the atoms, not for the bits and this is why we offer free learning through our open-source and open-hardware products... Our eagerness to learn is relentless. And we want to share it. Our own production line lets us work fast for customers who put the emphasis on quality and smart designs. Made in France stamped stuff !.