Más contenido relacionado La actualidad más candente (20) Similar a Powering Dynamic M2M Event Processing with OSGi - W Bowers (20) Powering Dynamic M2M Event Processing with OSGi - W Bowers1. © Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Powering Dynamic M2M Event
Processing with OSGi
Dynamic Complex Event Processing and
OSGi
Hitachi Communication Technologies America, Inc.
6/12/2014
Walt Bowers
Chief Architect OSGi Solutions
2. Contents
© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
1. The Vs of Big Data
2. Complex Event Processing
3. Dynamic Complex Event Processing
Dynamic Complex Event Processing and OSGi.
Powering Dynamic M2M Event Processing with OSGi
1
4. Demo
3. © Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
1. The Vs of Big Data
Dynamic Complex Event Processing and OSGi.
Powering Dynamic M2M Event Processing with OSGi
2
5. 4© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
The missing “V” of Big Data
Extracting VALUE from VIABLE Data
WHERE It Matters and WHEN It Matters
The missing “V” of Big Data
7. 6© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Key Elements of Connected Intelligence
Intelligent
Devices
Always-on devices
connected to variety
of sensors and
running multiple
software
applications
1
Real-Time
Analytics
High-frequency
data analysis for
instant decision
making and
automation of
information flows
2
Big Data
Integration of data
from connected
devices with
enterprise
applications and
historical data
3
8. 7© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamic Predictive Analytics
• Local analytics and business rules are controlled by
global analytics
– In-flight data analytics on the device
– Near real time response on the device
• Global Analytics for the Big Patterns
– Big Data post processing
– Discover Hidden Patterns/dependencies
• Dynamically Adjust the Rules
– Update new rules to the local device
– Enhances the devices local analytics
• Rinse and Repeat
9. 8© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Transportation Example
Passenger
Counter
M2M Gateway)
1 Location tracking (LBS)
Camera (Normal operation,
local storage; emergencies such
as Amber alert, streaming to
emergency authorities)
Digital signage (Information
and advertising - Location
aware)
Vehicle diagnostics
(Sensors throughout
vehicle)
Route monitoring (Time of
day optimization)
Fraud detection
(Passenger counter,
camera, fare collection)
3
2
4
56
7
10. © Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
2. Complex Event Processing
Dynamic Complex Event Processing and OSGi.
9
Powering Dynamic M2M Event Processing with OSGi
11. 10© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Intelligence Is Real-Time, Event-Based Analytics
Communication
Events
Machine
Events
Security
Events
Environmental
Events
Business
Logic
Events
Complex Events Processing
enables real-time business insights from edge devices
12. 11© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Complex Event Processing (CEP)
• Event-driven Architecture
• A generic data management infrastructure for
processing in-flight data before data is
potentially stored to deliver results in near
real-time
• Programming language for defining rules
• It allows users to
Aggregate/Correlate/Enrich/Detect Patterns
in high speed streaming data
13. 12© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Complex Event Processing (CEP)
Data/Msg.
Feeds
Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
CEP Application Container
14. 13© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Event Capture
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
CEP Application Container
• Events generated at sources
• Adaptor captures event and sends it into the
Event Processing Network
15. 14© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Event Processing
• Events processed in flight
• Merging multiple event sources and types
• Data enrichment by accessing external data sources
(e.g. databases)
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
CEP Application Container
16. 15© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Event Dispatch
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
CEP Application Container
• Processing produces events
• Adaptor receives event and sends it into the
downstream clients
17. © Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
3. Dynamic Complex Event Processing
Dynamic Complex Event Processing and OSGi.
16
Powering Dynamic M2M Event Processing with OSGi
18. 17© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamic Environment
• Devices do not operate in a static environment
• Inputs change
• Knowledge is gained from analytics
• Additional systems want to receive the output
19. 18© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamic Behavior
• Our Complex Event Processing engine needs
to be dynamic
• Ability to change behavior without stopping the
flow
• Allow a higher level system to change the
processing rules
20. 19© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Enter OSGi
Powered by
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
CEP Application Container
OSGi Framework
Java
• Dynamic Modular System for Java
• Mature Lightweight Application
Framework
– Ideal for embedded environments
• Supports Module Lifecycle
– Install/start/stop/uninstall/upgrade
– Remotely manageable
– Versioning
• Services Model
– Advertise and discover services
– Modules are dependent on service
not implementation
21. 20© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
OSGi Deployment Environment
OSGi Management System
&
Repository
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
OSGi - CEP
OSGi Framework
Java
Remote Device
22. 21© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamically Changing Behavior
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
OSGi CEP
OSGi Management System
&
Repository
Remote Device
Deployed System. Happily processing…
23. 22© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamically Changing Behavior
The Data Inputs Change
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
OSGi CEP
24. 23© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamically Changing Behavior
Update the adaptor
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
OSGi CEPOSGi Management System
&
Repository
Remote Device
25. 24© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamically Changing Behavior
Process the Events Differently
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
OSGi CEP
OSGi Management System
&
Repository
Remote Device
26. 25© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamically Changing Behavior
Forward to additional locations for processing
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
OSGi CEP
OSGi Management System
&
Repository
Remote Device
27. 26© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamically Changing Behavior
Happily processing again…
OSGi Management System
&
Repository
Remote Device
Data/Msg.
Feeds Listener/SINK:
User Code
(Plain Java)
Data Feed
Adapters
Process
Events
(CEP)
OSGi CEP
28. © Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
4. Demo
How OSGi and Java enables smart data on M2M aggregators and gateways.
Powering Dynamic M2M Event Processing with OSGi
27
30. 29© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
The Components
LCD
Display
Phidget
Temp
Sensor
31. 30© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
The Components
• OSGi based
• Continuous Query Language (CQL) for defining rules
• http://www.oracle.com/us/technologies/java/embedded/event-
processing/overview/index.html?ssSourceSiteId=opn
• Arm Based Linux platform
• http://www.raspberrypi.org/
• USB hardware devices
• Open and inexpensive
• http://www.phidgets.com/
• Hitachi’s OSGi Framework
32. 31© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamic Behavior In Action
Start reporting temperature changes below ambient temperature
33. 32© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamic Behavior In Action
Change the rules and redeploy remotely
34. 33© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Dynamic Behavior In Action
Now reporting temperature changes above ambient
temperature
35. © Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Hitachi Communication Technologies America, Inc.
Dynamic Complex Event Processing and OSGi.
Powering Dynamic M2M Event Processing with OSGi
6/12/2014
Walt Bowers
Chief Architect OSGi Solutions
Walt.bowers@hitachi-cta.com
END
34
39. 38© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Volume
3 006 477 107 200 GB added to the “digital universe”
~30% of it generated by machines
42 949 673 000 000 GB – 15 x increase
42 % will be generated by devices
Assuming
2GB/hr it will
take 2.3 billion
years to
download
According to IDC’s “Digital Universe in 2020“ study published in December
2012
41. 40© Hitachi Communication Technologies America, Inc. 2014. All rights reserved.
Variety
•Clickstream Data
•Twitter Feeds
•Facebook Postings
•Web Content
Web and Social Media
•Smart Meters Readings
•RFID Readings
•Oil Rig Sensors
•GP Signals
Machine-to-Machine
•Healthcare Claims
•Telecommunication Call Details Record
•Utility Billing Records
Big Transaction Data
•Facial Recognition
•Genetics
Biometrics
•Call Center Voice Recording
•Email
•Electronic Medical Records
Human Generated