Warp 10: the key enabler of digital transformation of IoT business - - PowerPoint PPT Presentation
Warp10 Warp 10: the key enabler of digital transformation of IoT business From sensors to industrial processes, from monitoring to cybersecurity, Warp10 proposes a disruptive software technology which makes you to consider data and artificial
From sensors to industrial processes, from monitoring to cybersecurity, Warp10 proposes a disruptive software technology which makes you to consider data and artificial intelligence in a new and different perspective for IoT business
Cityzen Data www.cityzendata.com / www.warp10.io
Warp10
Warp 10: A powerful and disruptive software platform to address sensors / IoT / Machine Data
#1
3
Machine Data Analytics
Four generic families of Data
#1
Data from major companies databases
Big Data = Data Mining No Disruption
#2
Data from Social Networks, mails …
Semantic, Content analytics
Relational SQL NoSQL
#3
Data for Sensors, Meters, Mobiles …
Where IoT Big Data is
NoSQL Time Series
#4
Geospatial databases
Structured Géospatial 3D Modeling
examples When traditional IT application is based on relational databases, IoT requires other tech and more flexible approaches
Warp10
4
Machine Data Analytics
5
Machine Data Analytics
From summary statistics to advanced signal processing and pattern detection
A stack based language dedicated to time series analytics
Things / Sensors Data transmission Data cleansing Data synchronisation Statistics, pattern detection, machine learning, correlations, anomalies detection … Data filtering Business Application Predictive
Warp10
6
Machine Data Analytics
Secured by Design
Metadata encrypted by default Total data encryption in option Dynamic allocation of cryptographic tokens Data access by token holders tracked Data portfolio management
7
Machine Data Analytics
Machine learning tasks:
behavior understanding
Tree-based Algorithms
Deep learning
(perceptron, convolutional, recurrent, long short term memory, ...)
networks
Statistical Learning
decision processes
Dimensionality reduction
embedding (t-SNE)
Warpscript proposes generic extensions to learn, to reproduce, to detect, to predict, or to simulate patterns, behaviors, outliers, anomalies, failures …
…or How to take profit of complex algorithms in business applications
8
Machine Data Analytics
HA Datalog
9
Machine Data Analytics
http://www.warp10.io
10
Machine Data Analytics
Cloud or On Premise IT Telecoms To an horizontal business value added for apps Sensors Meters IoT
Energy App Transport App Health App
Warp10 by Cityzen Data
A horizontal, neutral and industrial approach on the sensors and IoT market
Cityzen Data ingests and manages more than 250 Bn measures / day in SaaS mode (Sept 2016)
Cityzen Data works with major companies
When many companies claim they have IoTData platform, we can prove that we have the best and the most promising one.
GE, Airbus, Orange, Amazon, OVH , Cap Gemini …
Cityzen Data helps clients to get value from their data
Cityzen Data is a tech companies without any interference with client
Businesses functions, algorithms and services developed by clients remain their property.
100,000 to 1.5 million measures / sec / Core
11
Machine Data Analytics
Ranked #5
Warp10
Warp10 Use Cases #2
13
Machine Data Analytics
Transmission Microphasers Ingesting Renewable energies from small sources to bigger ones in real time Market / trading Industrial Consumption Cities, Neighborhoods, Buildings Storage Home Consumption
14
Machine Data Analytics
Future mobility is typically the use case that requires to cross a large range of data. With CIRB in Brussels, Cityzen Data aims to :
Multimodal mobility management based on gathering and crossing data from:
15
Machine Data Analytics
Smart Building and Smart Factories are controlled by a large range of sensors in different areas: electricity, lighting, heating, water, cooling, security …. Storing and crossing data allow to get value from a large amount of existing data
16
Machine Data Analytics
Source: Airbus
17
Machine Data Analytics
A Self-DrivingCar Will Create 1 Gigabyte of Data Per Second
https://datafloq.com/read/self- driving-cars-create-2-petabytes- data-annually/172
Last year, an estimated 26 million connected cars collected more than 480 Terabytes of data. That number is expected to increase to 11.1 petabytes by 2020. Some plug-in hybrid vehicles are capable
http://www.ibmbigdatahub.com/blog/how- vehicle-telematics-changing-auto-industry
http://www.economist.com/blogs/economist- explains/2013/04/economist-explains-how- self-driving-car-works-driverless
18
Machine Data Analytics
Data analytics:
America’s Cup Bermuda 2017
19
Machine Data Analytics
Warp 10 cross and analyzes data coming from any type of sensors, subsystems, scada … to detect fault, anomalies, discords or to predict future troubles Monitoring of systems is done by sensors measuring a large range of parameters: mechanic, power, data transmission, data processing, latency …
20
Machine Data Analytics
Data Analytics to prevent risks on the yields and market value of production Warp 10 leveraged to cross and analyze :
(manure spreaders, irrigation …)
21
Machine Data Analytics
The original use case for Cityzen Data : getting data from smart textile in order :
Movements Heart rate / ECG Temperature PH
22
Machine Data Analytics
Security and cybersecurity need to cross data from a large range of sources. Geo Time Series™ technologies is particularly relevant to detect weak signals among the ocean of various types of data. The Warp10 technology is used to monitor Internet accesses.
Internet access Internet surfing Smartphones Social media Personal connections Phone calls Mails Payments Hard disks contents Borders controls Airport controls Interview minutes Videosecurity
23
Machine Data Analytics
To get real value in maintenance, sensors data need: To be synchronized To be historicized To be cleaned To be modeled To be crossed with other data To be interpreted in real time (and in batch mode) Cityzen Data provides all these features through “on the shelves” tools & algorithms to different sectors: industry, power generators, cars, aeronautics, telecoms …
A strong trend: “Where Predictive Analytics Is Having the Biggest Impact”. Harvard Business Review
24
Machine Data Analytics
Testbeds are facing to the increasing of the number of physical sensors and systems
huge amount of measures. Warp10 provides “on the shelves” features and tools to testbeds experts to get benefit from:
Beyond the tech, a new paradigm in IT architectures #3
26
Machine Data Analytics
Business Process Assets HR Devices Sensors Systems IT architecture driven by transactions Rigid business applications & databases IoT Sensors Events Transactions IT architecture driven by stream
Flexible and scalable architecture Operations
Cityzen Data
27
Machine Data Analytics
Data are « owned » by a business or specific entity 1- Ownership Data refer to a specific format and ontology 2- Interoperability Third applications decrease primary applications performances 3- Load balancing « Non relevant » data are removed by primary applications 4- Missing Information SQL technologies have weak performance for IoT Data 5 - Performances Difficult to face to new and unknown future sources of data 8- Non flexibility Data management tied to vendors applications 7- Non independency SQL BI request heavy extraction operation (Datawarehouse) 6- Rigid extractions
28
Machine Data Analytics
Data close to row data are easier to be managed regarding ownership 1- Ownership Time series data are defined by a universal format 2- Interoperability « Time series » architectures are scalable 3- Load computing All data – even mistakes – are stored 4- Missing information Hig level performances for IoT Data (x10, x100) 5 - Performances Natural ingestion of new sources of data without any impact 8- Flexibility Neutral architecture regarding applications 7- Independency Dynamic accesses to selected data 6- On the fly accesses
29
Machine Data Analytics
Data is mainly considered as a property belonging to a specific unit within an organization
TODAY
Data result from one application when it is required to address new usages and services Development of new applications requires complex accesses to existing data or interfaces with existing data or applications Time Series can be a way to store raw data and favors an global asset approach
TOMORROW
Historian of raw data is the key to develop applications and services when needs occur New applications and services can be developed in a short time
06/12/2016 Cityzen Data
30
I am coming out from a Warp 10 @cityzendata pres … I am feeling like Howard Carter
it will no longer be the same
Mathias Herberts – CTO mathias.herberts@cityzendata.com Hervé Rannou – CEO herve.rannou@cityzendata.com